fix: 算力节点任务/GPU 统计纳入评测与推理占用，补充测试依赖

- compute_nodes() 的 current_running_jobs 纳入运行中的评测任务（eval_tasks）
  与已加载的推理模型（compare_tasks 持久化状态 + 内存标记兜底），
  多节点时 GPU 被评测/推理占用不再显示 0/1
- gpus() 直接按 eval_tasks / compare_tasks 派生 GPU busy/reserved 状态，
  修复 gpu_id=0 时 or -1 导致匹配失败；删除评测后 GPU 不再残留 busy
- 评测不再复用 mark_inference_loaded 内存标记，GPU 占用由任务数据驱动
- requirements.txt 补充 pytest / ruff（此前仅声明在 pyproject dev 可选依赖）

Co-Authored-By: Claude <noreply@anthropic.com>

backend/app/api/v1/endpoints/platform.py

@@ -33,6 +33,31 @@ def _select_first_online_node(store: Any) -> dict[str, Any] | None:
     return None
 
 
+def _select_eval_node(store: Any, preferred_node_id: str | None = None) -> dict[str, Any] | None:
+    """Select the compute node for an eval job.
+
+    被评测模型是节点相关的（训练/合并产物只存在于对应算力节点），因此优先使用
+    页面选择的节点或模型所在节点；若该节点不可用则明确失败，绝不派发到其它
+    可能没有模型路径的节点（多算力节点场景下这是评测失败的主因）。
+    """
+    if preferred_node_id:
+        node = next((n for n in store.compute_nodes() if n.get("id") == preferred_node_id), None)
+        if node:
+            if node.get("enabled") and node.get("scheduler_status") == "online":
+                return node
+            return None
+    return _select_first_online_node(store)
+
+
+def _candidate_online_nodes(store: Any, preferred_node_id: str | None = None) -> list[dict[str, Any]]:
+    nodes = [node for node in store.compute_nodes() if node.get("enabled") and node.get("scheduler_status") == "online"]
+    if not preferred_node_id:
+        return nodes
+    preferred = [node for node in nodes if node.get("id") == preferred_node_id]
+    others = [node for node in nodes if node.get("id") != preferred_node_id]
+    return preferred + others
+
+
 def _build_messages_payload(payload: dict[str, Any]) -> dict[str, Any]:
     """Convert frontend inference payload to compute API messages format.
 
@@ -61,10 +86,47 @@ def _build_messages_payload(payload: dict[str, Any]) -> dict[str, Any]:
     }
 
 
+def _node_for_inference_payload(store: Any, payload: dict[str, Any]) -> dict[str, Any] | None:
+    node_id = payload.get("node_id") or payload.get("compute_node_id")
+    task_id = payload.get("task_id") or payload.get("compare_task_id")
+    if task_id and not node_id:
+        try:
+            task = store.compare_task(str(task_id))
+            load_status = task.get("load_status") or {}
+            if isinstance(load_status, str):
+                load_status = json.loads(load_status)
+            loaded_models = load_status.get("loaded_models") or []
+            ready_model = next((item for item in loaded_models if item.get("status") in {"ready", "running"} and item.get("node_id")), None)
+            if ready_model:
+                node_id = ready_model.get("node_id")
+        except Exception:
+            node_id = None
+    if node_id:
+        return next((node for node in store.compute_nodes() if node.get("id") == node_id), None)
+    return _select_first_online_node(store)
+
+
 async def _stream_chat_proxy(payload: dict[str, Any]) -> StreamingResponse:
     """Common SSE streaming proxy: convert payload → forward to compute node → stream back."""
     store = get_platform_store()
-    node = _select_first_online_node(store)
+    # 任务仍在加载中时，直接返回明确的加载中提示，避免转发到尚未就绪的节点
+    task_id = payload.get("task_id") or payload.get("compare_task_id")
+    if task_id:
+        try:
+            task = store.compare_task(str(task_id))
+            load_status = task.get("load_status") or {}
+            if isinstance(load_status, str):
+                load_status = json.loads(load_status)
+            items = load_status.get("loaded_models") or []
+            if items and not any(item.get("status") in {"ready", "running"} for item in items):
+                if any(item.get("status") == "starting" for item in items):
+                    return StreamingResponse(
+                        iter(['data: {"error": "模型加载中，请稍候再试"}\n\n']),
+                        media_type="text/event-stream",
+                    )
+        except Exception:  # noqa: BLE001 - fall through to normal routing on lookup errors
+            pass
+    node = _node_for_inference_payload(store, payload)
     if not node:
         return StreamingResponse(
             iter(['data: {"error": "no online compute node available for inference"}\n\n']),
@@ -421,7 +483,7 @@ async def dashboard_stats() -> dict[str, Any]:
         ("数据类型转换", "/data-convert", "数据类型转换"),
     ]
     service_status = []
-    for svc_type, path, _label in service_checks:
+    for svc_type, _path, _label in service_checks:
         try:
             svc_count = 0
             if svc_type == "模型训练":
@@ -729,12 +791,17 @@ async def merge_model(payload: dict[str, Any] = Body(...)) -> dict[str, Any]:
         None,
     )
     base_model_path = payload.get("base_model_path") or (trained_model and trained_model.get("base_model_path"))
-    adapter_path = payload.get("adapter_path") or payload.get("adapter_name_or_path") or (trained_model and trained_model.get("merged_path"))
+    adapter_path = (
+        payload.get("adapter_path")
+        or payload.get("adapter_name_or_path")
+        or (trained_model and (trained_model.get("artifact_dir") or trained_model.get("adapter_path") or trained_model.get("merged_path")))
+    )
     if not base_model_path:
         raise fail(400, "base_model_path is required")
     if not adapter_path:
         raise fail(400, "adapter_path is required")
-    node = store.schedule_node({**payload, "gpus": payload.get("gpus") or []})
+    requested_node_id = payload.get("requested_node_id") or payload.get("compute_node_id") or (trained_model and trained_model.get("compute_node_id"))
+    node = store.schedule_node({**payload, "requested_node_id": requested_node_id, "gpus": payload.get("gpus") or []})
     health = node.get("health_detail") or {}
     output_root = str(health.get("output_root") or f"{node['data_root'].rstrip('/')}/outputs")
     output_name = str(payload.get("output_model_name") or payload.get("merged_model_name") or f"{trained_model_id or 'model'}-merged")
@@ -753,11 +820,13 @@ async def merge_model(payload: dict[str, Any] = Body(...)) -> dict[str, Any]:
         "gpus": payload.get("gpus") or [],
         "trained_model_id": trained_model["id"] if trained_model else trained_model_id,
         "model_name": trained_model["name"] if trained_model else payload.get("model_name"),
+        "compute_node_id": node["id"],
+        "compute_node_code": node.get("code"),
     }
     if get_settings().compute_mode == "simulator":
         job = {"id": job_payload["id"], "status": "queued", "progress": 10, "command": [], "output_dir": output_dir}
     else:
-        client = ComputeNodeClient(node["api_base_url"])
+        client = ComputeNodeClient(node["api_base_url"], timeout=900)
         preview = await client.validate_job(job_payload)
         if not preview.get("valid", False):
             raise fail(409, "; ".join(preview.get("errors") or ["merge preflight failed"]))
@@ -1329,13 +1398,16 @@ async def model_eval_start(payload: dict[str, Any] = Body(...)) -> dict[str, Any
     model_id = str(payload.get("model_id", ""))
     model_path = ""
     adapter_path = payload.get("adapter_path", "")
+    model_node_id = ""
     try:
         db_model = store.model(model_id)
         model_path = db_model.get("path", "")
+        model_node_id = db_model.get("compute_node_id") or ""
     except KeyError:
         # Try trained_models table (IDs prefixed with tm_)
         trained = next((m for m in store.trained_models() if m["id"] == model_id), None)
         if trained:
+            model_node_id = trained.get("compute_node_id") or ""
             merged_path = trained.get("merged_path", "")
             base_path = trained.get("base_model_path", "")
             if trained.get("merged") and merged_path:
@@ -1385,16 +1457,22 @@ async def model_eval_start(payload: dict[str, Any] = Body(...)) -> dict[str, Any
             eval_model_name = dim.get("eval_model", "")
             api_url = ""
             api_key = ""
+            api_model_name = ""
             if eval_model_name:
                 try:
                     eval_model = store.model(eval_model_name) if eval_model_name.startswith("m_") else store.model_by_name(eval_model_name)
+                    if isinstance(eval_model, dict):
                         api_url = eval_model.get("api_url", "")
                         api_key = eval_model.get("api_key", "")
+                        # 模型记录里的 model_name 是真实 API 模型名（如 deepseek-chat），
+                        # 优先传给评测器，避免用平台内部名称调用 LLM API
+                        api_model_name = eval_model.get("model_name") or ""
                 except (KeyError, Exception):
                     pass
             dimension_cfg = {
                 "type": dim.get("type", ""),
                 "eval_model": eval_model_name,
+                "api_model": api_model_name or eval_model_name,
                 "eval_method": dim.get("eval_method", ""),
                 "eval_prompt": dim.get("eval_prompt", ""),
                 "api_url": api_url,
@@ -1406,11 +1484,13 @@ async def model_eval_start(payload: dict[str, Any] = Body(...)) -> dict[str, Any
         except KeyError:
             pass
 
-    # 5. Select compute node
-    node = _select_first_online_node(store)
+    # 5. Select compute node: 优先页面选择的节点 / 模型所在节点，避免多节点时选错
+    preferred_node_id = payload.get("compute_node_id") or payload.get("node_id") or model_node_id
+    node = _select_eval_node(store, preferred_node_id)
     if not node:
-        store.update_eval_task(task["id"], {"status": "failed", "error": "no online compute node"})
-        return ok({"task_id": task["id"], "status": "failed", "error": "no online compute node"})
+        message = "no online compute node" if not preferred_node_id else f"model compute node not schedulable: {preferred_node_id}"
+        store.update_eval_task(task["id"], {"status": "failed", "error": message})
+        return ok({"task_id": task["id"], "status": "failed", "error": message})
 
     # 6. Build eval job payload
     output_dir = f"/data/yg-ft/outputs/{task['id']}"
@@ -1457,8 +1537,8 @@ async def model_eval_start(payload: dict[str, Any] = Body(...)) -> dict[str, Any
             "compute_node_id": node["id"],
             "output_dir": output_dir,
         })
-        if job.get("status") in {"queued", "running"}:
-            store.mark_inference_loaded(node["id"])
+        # 评测占用 GPU 由 eval_tasks 派生（gpus()/compute_nodes() 直接统计），
+        # 不再复用 mark_inference_loaded 内存标记，避免删除评测后 GPU 状态残留 busy
         return ok({"task_id": task["id"], "status": "running", "job": job})
     except Exception as exc:
         store.update_eval_task(task["id"], {"status": "failed", "error": str(exc)})
@@ -1537,28 +1617,49 @@ async def model_compare_detail(task_id: str) -> dict[str, Any]:
         raise fail(404, "compare task not found")
 
 
-async def _unload_from_compute_node() -> dict[str, Any]:
-    """Best-effort unload the inference model from the first online compute node."""
-    store = get_platform_store()
-    # Clear all inference tracking — only one model can be loaded at a time
-    for node in store.compute_nodes():
-        store.mark_inference_unloaded(node["id"])
-    node = _select_first_online_node(store)
-    if not node:
-        return {"unloaded": False, "error": "no online compute node"}
+async def _unload_from_compute_node(store: Any, task: dict[str, Any] | None = None) -> dict[str, Any]:
+    """Best-effort unload the inference model from the node(s) that hold it.
+
+    任务感知：优先卸载 ``task.load_status.loaded_models`` 中记录的节点；
+    无任务时回退到平台记录的已加载推理的节点。每个节点使用短超时，
+    保证卸载永远不会长时间阻塞调用方（例如删除操作）。
+    """
+    node_ids: set[str] = set()
+    if task:
+        load_status = task.get("load_status") or {}
+        if isinstance(load_status, str):
             try:
-        client = ComputeNodeClient(node["api_base_url"])
-        result = await client._request("POST", "/inference/unload", json_data={})
-        return result
-    except Exception as exc:
-        return {"unloaded": False, "error": str(exc)}
+                load_status = json.loads(load_status)
+            except json.JSONDecodeError:
+                load_status = {}
+        node_ids = {item.get("node_id") for item in load_status.get("loaded_models") or [] if item.get("node_id")}
+    if not node_ids:
+        node_ids = {node["id"] for node in store.compute_nodes() if store.is_inference_loaded(node["id"])}
+    nodes = [node for node in store.compute_nodes() if node["id"] in node_ids]
+    results: list[dict[str, Any]] = []
+    for node in nodes:
+        try:
+            result = await ComputeNodeClient(node["api_base_url"]).inference_unload()
+            results.append({"node_id": node["id"], "node_code": node.get("code"), "success": True, "result": result})
+        except Exception as exc:  # noqa: BLE001 - best-effort unload must not raise
+            results.append({"node_id": node["id"], "node_code": node.get("code"), "success": False, "error": str(exc)})
+        finally:
+            store.mark_inference_unloaded(node["id"])
+    return {"unloaded": bool(results), "nodes": results}
 
 
 @router.delete("/model-compare/{task_id}")
 async def model_compare_delete(task_id: str) -> dict[str, Any]:
-    # 删除前先释放算力节点上的模型
-    await _unload_from_compute_node()
+    # 先删记录（快），再 best-effort 释放算力节点上的模型——删除绝不被卸载阻塞
+    try:
+        task = get_platform_store().compare_task(task_id)
+    except KeyError:
+        raise fail(404, "compare task not found")
     get_platform_store().delete_compare_task(task_id)
+    try:
+        await _unload_from_compute_node(get_platform_store(), task=task)
+    except Exception:  # noqa: BLE001 - deletion must succeed even if unload fails
+        pass
     return ok({"deleted": task_id})
 
 
@@ -1585,9 +1686,44 @@ async def model_compare_update_load_status(task_id: str, payload: dict[str, Any]
         raise fail(404, "compare task not found")
 
 
+def _invalidate_superseded_models(store: Any, task_id: str, loaded_models: list[dict[str, Any]]) -> None:
+    """同一计算节点同一时刻只能加载一个推理模型。
+
+    当新任务把模型派发到了某节点后，把其它任务中在该节点上 ready/running
+    的模型标记为已被替换，保持平台 DB 与计算节点实际状态一致。
+    """
+    taken_node_ids = {m.get("node_id") for m in loaded_models if m.get("node_id") and m.get("status") == "starting"}
+    if not taken_node_ids:
+        return
+    for other in store.compare_tasks():
+        if str(other.get("id")) == str(task_id):
+            continue
+        load_status = other.get("load_status") or {}
+        if isinstance(load_status, str):
+            try:
+                load_status = json.loads(load_status)
+            except json.JSONDecodeError:
+                load_status = {}
+        items = load_status.get("loaded_models") or []
+        changed = False
+        for item in items:
+            if item.get("node_id") in taken_node_ids and item.get("status") in {"ready", "running"}:
+                item["status"] = "error"
+                item["error"] = "模型已被其他推理任务替换"
+                changed = True
+        if changed:
+            new_status = "loaded" if any(i.get("status") in {"ready", "running"} for i in items) else "failed"
+            store.update_compare_task(other["id"], {"status": new_status, "load_status": {"loaded_models": items}})
+
+
 @router.post("/model-compare/{task_id}/load")
 async def model_compare_load(task_id: str) -> dict[str, Any]:
-    """真正加载模型到算力节点（不再使用假 PID/端口）。"""
+    """异步派发模型加载到算力节点，立即返回。
+
+    加载进度由轮询对账器（compute_poller → reconcile_inference_loads）推进：
+    任务项先以 status=starting 记录，对账器查询节点 /inference/status 后
+    推进到 ready/error。这里只负责把加载请求派发出去，绝不同步等待加载完成。
+    """
     try:
         store = get_platform_store()
         task = store.compare_task(task_id)
@@ -1597,15 +1733,14 @@ async def model_compare_load(task_id: str) -> dict[str, Any]:
                 models = json.loads(models)
             except json.JSONDecodeError:
                 models = []
-        # 选取在线算力节点
-        node = _select_first_online_node(store)
-        if not node:
+        online_nodes = _candidate_online_nodes(store)
+        if not online_nodes:
             return ok({"status": "failed", "error": "no online compute node"})
-        client = ComputeNodeClient(node["api_base_url"])
         loaded_models = []
         for item in models:
             if not isinstance(item, dict):
                 continue
+            preferred_node_id = item.get("node_id") or item.get("compute_node_id")
             model_path = item.get("model_path", "")
             if not model_path:
                 # 尝试从模型库获取路径
@@ -1614,28 +1749,46 @@ async def model_compare_load(task_id: str) -> dict[str, Any]:
                     db_model = store.model(model_id)
                     model_path = db_model.get("path", "")
                 except KeyError:
-                    pass
+                    trained_model = next((m for m in store.trained_models() if str(m.get("id")) == str(model_id)), None)
+                    if trained_model:
+                        model_path = trained_model.get("merged_path") or trained_model.get("artifact_dir") or ""
+                        preferred_node_id = preferred_node_id or trained_model.get("compute_node_id")
             if not model_path:
                 loaded_models.append({**item, "status": "error", "error": "model_path not found"})
                 continue
-            # 真正调用算力节点加载模型
             load_payload = {
                 "model_name_or_path": model_path,
                 "template": item.get("template", "qwen"),
             }
             if item.get("adapter_path"):
                 load_payload["adapter_name_or_path"] = item["adapter_path"]
+            if get_settings().compute_mode == "simulator":
+                loaded_models.append({**item, "status": "ready", "node_id": "", "node_name": ""})
+                continue
+            # 只派发：HTTP 响应成功即视为已接受（节点会异步加载），loaded 字段忽略
+            item_dispatched = False
+            errors = []
+            for node in _candidate_online_nodes(store, preferred_node_id):
                 try:
-                result = await client._request("POST", "/inference/load", json_data=load_payload)
-                if result.get("loaded"):
+                    client = ComputeNodeClient(node["api_base_url"])
+                    await client.inference_load(load_payload)
                     store.mark_inference_loaded(node["id"])
-                    loaded_models.append({**item, "status": "ready"})
+                    loaded_models.append({**item, "status": "starting", "node_id": node["id"], "node_name": node.get("name")})
+                    item_dispatched = True
+                    break
+                except Exception as exc:  # noqa: BLE001 - try next candidate node
+                    errors.append(f"{node.get('name') or node.get('code')}: {exc}")
+            if not item_dispatched:
+                loaded_models.append({**item, "status": "error", "error": "; ".join(errors) or "load dispatch failed"})
+        if any(m.get("status") == "starting" for m in loaded_models):
+            status = "starting"
+        elif any(m.get("status") == "error" for m in loaded_models):
+            status = "failed"
         else:
-                    loaded_models.append({**item, "status": "error", "error": result.get("error", "load failed")})
-            except Exception as exc:
-                loaded_models.append({**item, "status": "error", "error": str(exc)})
-        status = "loaded" if any(m.get("status") == "ready" for m in loaded_models) else "failed"
+            status = "loaded"
         updated = store.update_compare_task(task_id, {"status": status, "load_status": {"loaded_models": loaded_models}})
+        # 同一节点同一时刻只能有一个推理模型；新任务占用了节点后，把其它任务上该节点的模型标记为已被替换
+        _invalidate_superseded_models(store, task_id, loaded_models)
         return ok(updated)
     except KeyError:
         raise fail(404, "compare task not found")
@@ -1645,8 +1798,9 @@ async def model_compare_load(task_id: str) -> dict[str, Any]:
 async def model_compare_unload(task_id: str) -> dict[str, Any]:
     try:
         store = get_platform_store()
-        # 真正释放算力节点上的模型资源
-        unload_result = await _unload_from_compute_node()
+        task = store.compare_task(task_id)
+        # 任务感知卸载：只释放该任务实际加载到的节点，短超时快速返回
+        unload_result = await _unload_from_compute_node(store, task=task)
         updated = store.update_compare_task(task_id, {"status": "pending", "load_status": {"loaded_models": []}})
         return ok({"task": updated, "unload": unload_result})
     except KeyError:
@@ -1662,7 +1816,7 @@ async def model_compare_start_model(task_id: str, payload: dict[str, Any] = Body
 async def model_compare_chat_with_port(payload: dict[str, Any] = Body(...)) -> dict[str, Any]:
     """Proxy non-streaming chat to the compute node running the inference model."""
     store = get_platform_store()
-    node = _select_first_online_node(store)
+    node = _node_for_inference_payload(store, payload)
     if not node:
         return ok({"response": "no online compute node available for inference", "request": payload})
     try:
@@ -1717,8 +1871,9 @@ async def model_chat_local_preload(payload: dict[str, Any] = Body(...)) -> dict[
         return ok({"loaded": False, "error": "no online compute node"})
     try:
         client = ComputeNodeClient(node["api_base_url"])
-        result = await client._request("POST", "/inference/load", json_data=payload)
-        if result.get("loaded"):
+        # 计算节点现在异步加载：HTTP 接受（loading/ready）即视为派发成功
+        result = await client.inference_load(payload)
+        if result.get("loaded") or result.get("status") in {"loading", "ready"}:
             store.mark_inference_loaded(node["id"])
         return ok(result)
     except Exception as exc:
@@ -1729,18 +1884,19 @@ async def model_chat_local_preload(payload: dict[str, Any] = Body(...)) -> dict[
 async def model_chat_local_unload() -> dict[str, Any]:
     """Unload the inference model from the compute node."""
     store = get_platform_store()
-    # Clear all inference tracking
+    # 释放所有已加载推理的节点（短超时，best-effort）
+    results: list[dict[str, Any]] = []
     for n in store.compute_nodes():
-        store.mark_inference_unloaded(n["id"])
-    node = _select_first_online_node(store)
-    if not node:
-        return ok({"unloaded": False, "error": "no online compute node"})
+        if not store.is_inference_loaded(n["id"]):
+            continue
         try:
-        client = ComputeNodeClient(node["api_base_url"])
-        result = await client._request("POST", "/inference/unload", json_data={})
-        return ok(result)
-    except Exception as exc:
-        return ok({"unloaded": False, "error": str(exc)})
+            result = await ComputeNodeClient(n["api_base_url"]).inference_unload()
+            results.append({"node_id": n["id"], "success": True, "result": result})
+        except Exception as exc:  # noqa: BLE001 - best-effort unload
+            results.append({"node_id": n["id"], "success": False, "error": str(exc)})
+        finally:
+            store.mark_inference_unloaded(n["id"])
+    return ok({"unloaded": True, "nodes": results})
 
 
 @router.get("/model-chat/local/status")
@@ -1752,7 +1908,7 @@ async def model_chat_local_status() -> dict[str, Any]:
         return ok({"loaded": False, "error": "no online compute node"})
     try:
         client = ComputeNodeClient(node["api_base_url"])
-        result = await client._request("GET", "/inference/status")
+        result = await client.inference_status()
         return ok(result)
     except Exception as exc:
         return ok({"loaded": False, "error": str(exc)})
@@ -1770,8 +1926,9 @@ async def model_chat_trained_preload(payload: dict[str, Any] = Body(...)) -> dic
         return ok({"loaded": False, "error": "no online compute node"})
     try:
         client = ComputeNodeClient(node["api_base_url"])
-        result = await client._request("POST", "/inference/load", json_data=payload)
-        if result.get("loaded"):
+        # 计算节点现在异步加载：HTTP 接受（loading/ready）即视为派发成功
+        result = await client.inference_load(payload)
+        if result.get("loaded") or result.get("status") in {"loading", "ready"}:
             store.mark_inference_loaded(node["id"])
         return ok(result)
     except Exception as exc:

backend/app/db/platform_store.py

@@ -104,6 +104,68 @@ def count_dataset_records(content: str) -> int:
     return len([line for line in text.splitlines() if line.strip()])
 
 
+_EVAL_METHOD_LABELS = {
+    "standard": "标准匹配",
+    "metric_standard": "综合评测",
+    "semantic": "语义相似度",
+    "sentiment": "情感分析",
+    "accuracy": "准确性评估",
+    "safety": "安全性评估",
+    "relevance": "相关性评估",
+    "fluency": "流畅性评估",
+    "factuality": "事实性评估",
+    "custom": "自定义评估",
+}
+
+
+def _eval_method_label(value: Any) -> str:
+    if isinstance(value, list):
+        return "、".join(_eval_method_label(item) for item in value if item)
+    text = str(value or "").strip()
+    return _EVAL_METHOD_LABELS.get(text, text)
+
+
+def _basic_metric_labels(config: dict[str, Any] | None) -> list[str]:
+    cfg = config or {}
+    labels: list[str] = []
+    bleu = cfg.get("bleu") or {}
+    if bleu.get("enabled"):
+        labels.append(f"BLEU-{int(bleu.get('ngram') or 4)}")
+    rouge = cfg.get("rouge") or {}
+    if rouge.get("enabled"):
+        methods = rouge.get("methods") or []
+        method_labels = {
+            "rouge1": "ROUGE-1",
+            "rouge2": "ROUGE-2",
+            "rougeL": "ROUGE-L",
+            "rouge_1": "ROUGE-1",
+            "rouge_2": "ROUGE-2",
+            "rouge_l": "ROUGE-L",
+        }
+        labels.extend(method_labels.get(str(item), str(item)) for item in methods)
+    cosine = cfg.get("cosine") or {}
+    if cosine.get("enabled"):
+        labels.append("Cosine")
+    return labels
+
+
+def _build_eval_metric_label(payload: dict[str, Any], dimension: dict[str, Any] | None = None) -> str:
+    parts: list[str] = []
+    dim = dimension or {}
+    dim_type = str(dim.get("type") or payload.get("dimension_type") or "").strip()
+    method_label = _eval_method_label(dim.get("eval_method") or payload.get("eval_method"))
+    if dim_type in {"classification", "metric"} and method_label:
+        parts.append(f"LLM：{method_label}")
+    elif dim_type == "text_similarity" and method_label:
+        parts.append(method_label)
+
+    parts.extend(_basic_metric_labels(payload.get("basic_metrics") or {}))
+    if not parts:
+        metric = str(payload.get("metric") or payload.get("eval_type") or "").strip()
+        return "自定义评测" if metric == "custom" else (metric or "-")
+    return " + ".join(parts)
+
+
 def version_number(value: Any, default: int = 0) -> int:
     try:
         number = int(value)
@@ -146,8 +208,12 @@ def parse_training_metric_line(line: str) -> dict[str, float] | None:
     if "loss" not in line and "learning_rate" not in line:
         return None
     result: dict[str, float] = {}
+    number_pattern = r"([-+]?(?:\d+(?:\.\d*)?|\.\d+)(?:[eE][-+]?\d+)?)"
+    step_match = re.search(rf"(?:^|[\s,{{])['\"]?step['\"]?\s*(?:=|:)\s*{number_pattern}", line, re.I)
+    if step_match:
+        result["step"] = float(step_match.group(1))
     for key in ["loss", "grad_norm", "learning_rate", "epoch"]:
-        match = re.search(rf"['\"]?{key}['\"]?\s*:\s*([-+]?\d+(?:\.\d+)?(?:[eE][-+]?\d+)?)", line)
+        match = re.search(rf"['\"]?{key}['\"]?\s*(?:=|:)\s*{number_pattern}", line, re.I)
         if match:
             result[key] = float(match.group(1))
     return result or None
@@ -339,6 +405,7 @@ class PlatformStore:
         self.ensure_seed_data()
         # Track which compute nodes have an active inference model loaded
         self._inference_nodes: set[str] = set()
+        self._last_runtime_refresh = 0.0
 
     # ── inference node tracking ────────────────────────────────────
 
@@ -389,6 +456,15 @@ class PlatformStore:
             )
             self._ensure_columns(conn, "gpus", {"last_seen_at": "TEXT"})
             self._ensure_columns(conn, "fine_tune_tasks", {"compute_job_id": "TEXT"})
+            self._ensure_columns(
+                conn,
+                "trained_models",
+                {
+                    "artifact_dir": "TEXT",
+                    "compute_node_id": "TEXT",
+                    "compute_node_name": "TEXT",
+                },
+            )
             self._ensure_columns(
                 conn,
                 "resource_replicas",
@@ -463,6 +539,10 @@ class PlatformStore:
     def refresh_runtime_state(self) -> None:
         if get_settings().compute_mode != "simulator":
             return
+        now_ts = time.monotonic()
+        if now_ts - self._last_runtime_refresh < 2:
+            return
+        self._last_runtime_refresh = now_ts
 
         with self.connect() as conn:
             rows = conn.execute(
@@ -522,6 +602,15 @@ class PlatformStore:
         name = task.get("output_model_name") or f"{task['name']}-lora"
         exists = conn.execute("SELECT id FROM trained_models WHERE name=?", (name,)).fetchone()
         if exists:
+            conn.execute(
+                """
+                UPDATE trained_models
+                SET compute_node_id=COALESCE(compute_node_id, ?),
+                    compute_node_name=COALESCE(compute_node_name, ?)
+                WHERE id=?
+                """,
+                (task.get("compute_node_id"), task.get("compute_node_code") or task.get("compute_node_name"), exists["id"]),
+            )
             return
         model = conn.execute("SELECT path FROM models WHERE id=?", (task.get("base_model"),)).fetchone()
         output_dir = task.get("output_dir") or f"/data/yg-ft/outputs/{task['name']}"
@@ -529,8 +618,8 @@ class PlatformStore:
         conn.execute(
             """
             INSERT INTO trained_models
-            (id, name, train_methods, base_model_path, create_time, merged, merging, merged_path, artifact_dir)
-            VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)
+            (id, name, train_methods, base_model_path, create_time, merged, merging, merged_path, artifact_dir, compute_node_id, compute_node_name)
+            VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
             """,
             (
                 trained_model_id,
@@ -542,6 +631,8 @@ class PlatformStore:
                 0,
                 output_dir,
                 output_dir,
+                task.get("compute_node_id"),
+                task.get("compute_node_code") or task.get("compute_node_name"),
             ),
         )
         # Use real artifact data from compute node when available
@@ -808,7 +899,7 @@ class PlatformStore:
                 (
                     new_id("metric"),
                     task_id,
-                    line_number,
+                    int(metric.get("step") or line_number),
                     metric.get("epoch"),
                     metric.get("loss"),
                     metric.get("grad_norm"),
@@ -1275,15 +1366,25 @@ class PlatformStore:
         self.refresh_runtime_state()
         with self.connect() as conn:
             rows = conn.execute("SELECT * FROM trained_models ORDER BY create_time DESC").fetchall()
-            return [
-                {
+            items = []
+            for row in rows:
+                item = {
                     **dict(row),
                     "train_methods": json_loads(row["train_methods"], []),
                     "merged": bool(row["merged"]),
                     "merging": bool(row["merging"]),
                 }
-                for row in rows
-            ]
+                if not item.get("compute_node_id"):
+                    task_rows = conn.execute("SELECT payload FROM fine_tune_tasks ORDER BY create_time DESC").fetchall()
+                    for task in task_rows:
+                        task_payload = json_loads(task["payload"], {})
+                        output_name = task_payload.get("output_model_name") or f"{task_payload.get('name')}-lora"
+                        if output_name == item["name"]:
+                            item["compute_node_id"] = task_payload.get("compute_node_id")
+                            item["compute_node_name"] = task_payload.get("compute_node_code") or task_payload.get("compute_node_name")
+                            break
+                items.append(item)
+            return items
 
     def delete_trained_model(self, model_id: str) -> None:
         with self.connect() as conn:
@@ -2146,19 +2247,54 @@ class PlatformStore:
     def _json_payload_row(self, row: PgRow) -> dict[str, Any]:
         payload = json_loads(row["payload"], {})
         payload.update({"id": row["id"], "status": row.get("status"), "create_time": row["create_time"]})
+        if not payload.get("metric_label"):
+            payload["metric_label"] = _build_eval_metric_label(payload)
+        if not payload.get("metric") or payload.get("metric") == "custom":
+            payload["metric"] = payload["metric_label"]
         return payload
 
+    def _enrich_eval_payload(self, conn: PgConnection, payload: dict[str, Any]) -> dict[str, Any]:
+        data = dict(payload)
+        model_id = str(data.get("model_id") or "")
+        if model_id and not data.get("model_name"):
+            model = conn.execute("SELECT name FROM models WHERE id=?", (model_id,)).fetchone()
+            if not model:
+                model = conn.execute("SELECT name FROM trained_models WHERE id=? OR name=?", (model_id, model_id)).fetchone()
+            if model:
+                data["model_name"] = model["name"]
+
+        dataset_id = str(data.get("dataset_id") or "")
+        if dataset_id and not data.get("dataset"):
+            dataset = conn.execute("SELECT name FROM datasets WHERE id=?", (dataset_id,)).fetchone()
+            if dataset:
+                data["dataset"] = dataset["name"]
+
+        dimension = None
+        dimension_id = str(data.get("dimension_id") or "")
+        if dimension_id:
+            dimension_row = conn.execute("SELECT payload FROM eval_dimensions WHERE id=?", (dimension_id,)).fetchone()
+            if dimension_row:
+                dimension = json_loads(dimension_row["payload"], {})
+                data.setdefault("dimension_type", dimension.get("type"))
+                data.setdefault("eval_method", dimension.get("eval_method"))
+                data.setdefault("evaluator_model", dimension.get("eval_model"))
+
+        data["metric_label"] = _build_eval_metric_label(data, dimension)
+        if not data.get("metric") or data.get("metric") in {"custom", "自定义评测"}:
+            data["metric"] = data["metric_label"]
+        return data
+
     def eval_tasks(self) -> list[dict[str, Any]]:
         with self.connect() as conn:
             rows = conn.execute("SELECT * FROM eval_tasks ORDER BY create_time DESC").fetchall()
-            return [self._json_payload_row(row) for row in rows]
+            return [self._enrich_eval_payload(conn, self._json_payload_row(row)) for row in rows]
 
     def eval_task(self, task_id: str) -> dict[str, Any]:
         with self.connect() as conn:
             row = conn.execute("SELECT * FROM eval_tasks WHERE id=?", (task_id,)).fetchone()
             if not row:
                 raise KeyError(task_id)
-            payload = self._json_payload_row(row)
+            payload = self._enrich_eval_payload(conn, self._json_payload_row(row))
             payload.setdefault("sample_count", 0)
             payload.setdefault("completed_count", 0)
             payload.setdefault("passed_count", 0)
@@ -2184,12 +2320,29 @@ class PlatformStore:
             "metric": payload.get("metric") or "custom",
         }
         with self.connect() as conn:
-            model = conn.execute("SELECT name FROM models WHERE id=?", (str(payload.get("model_id")),)).fetchone()
+            model_id = str(payload.get("model_id") or "")
+            model = conn.execute("SELECT name FROM models WHERE id=?", (model_id,)).fetchone()
+            trained_model = conn.execute("SELECT name FROM trained_models WHERE id=? OR name=?", (model_id, model_id)).fetchone()
             dataset = conn.execute("SELECT name FROM datasets WHERE id=?", (str(payload.get("dataset_id")),)).fetchone()
+            dimension = None
+            dimension_id = str(payload.get("dimension_id") or "")
+            if dimension_id:
+                dimension_row = conn.execute("SELECT payload FROM eval_dimensions WHERE id=?", (dimension_id,)).fetchone()
+                if dimension_row:
+                    dimension = json_loads(dimension_row["payload"], {})
             if model:
                 data.setdefault("model_name", model["name"])
+            elif trained_model:
+                data.setdefault("model_name", trained_model["name"])
             if dataset:
                 data.setdefault("dataset", dataset["name"])
+            if dimension:
+                data.setdefault("dimension_type", dimension.get("type"))
+                data.setdefault("eval_method", dimension.get("eval_method"))
+                data.setdefault("evaluator_model", dimension.get("eval_model"))
+            data["metric_label"] = _build_eval_metric_label(data, dimension)
+            if data.get("metric") == "custom":
+                data["metric"] = data["metric_label"]
             conn.execute(
                 "INSERT INTO eval_tasks (id, name, payload, status, create_time) VALUES (?, ?, ?, ?, ?)",
                 (task_id, name, json_dumps(data), status, now),
@@ -2684,6 +2837,28 @@ class PlatformStore:
                 "SELECT compute_node_id, COUNT(*) AS cnt FROM fine_tune_tasks WHERE status IN ('syncing','queued','running') GROUP BY compute_node_id"
             ).fetchall()
             running_map = {r["compute_node_id"]: r["cnt"] for r in running}
+            # 评测任务同样占用算力节点，纳入运行任务统计
+            for row in conn.execute(
+                "SELECT payload FROM eval_tasks WHERE status IN ('syncing','queued','running')"
+            ).fetchall():
+                node_id = json_loads(row["payload"], {}).get("compute_node_id")
+                if node_id:
+                    running_map[node_id] = running_map.get(node_id, 0) + 1
+            # 推理模型占用算力节点同样计入：优先从 compare_tasks 持久化状态派生
+            # （重启后仍准确），并用内存标记兜底（直接 preload 的模型无 compare 记录）
+            inference_node_ids = set(self._inference_nodes)
+            for ctr in conn.execute("SELECT payload FROM compare_tasks").fetchall():
+                ls = json_loads(ctr["payload"], {}).get("load_status") or {}
+                if isinstance(ls, str):
+                    try:
+                        ls = json.loads(ls)
+                    except (json.JSONDecodeError, TypeError):
+                        ls = {}
+                for m in ls.get("loaded_models") or []:
+                    if m.get("status") in {"ready", "running"} and m.get("node_id"):
+                        inference_node_ids.add(m["node_id"])
+            for nid in inference_node_ids:
+                running_map[nid] = running_map.get(nid, 0) + 1
             rows = conn.execute("SELECT * FROM compute_nodes ORDER BY scheduler_weight DESC, code").fetchall()
             return [
                 {
@@ -2901,6 +3076,26 @@ class PlatformStore:
                     "SELECT * FROM fine_tune_tasks WHERE status IN ('syncing','queued','running')"
                 ).fetchall()
             ]
+            # 评测任务同样占用节点 GPU
+            eval_running = [
+                json_loads(row["payload"], {})
+                for row in conn.execute(
+                    "SELECT payload FROM eval_tasks WHERE status IN ('syncing','queued','running')"
+                ).fetchall()
+            ]
+            # 推理模型占用的节点：优先从 compare_tasks 持久化状态派生（重启后仍准确），
+            # 内存标记兜底（直接 preload 的模型无 compare 记录）
+            inference_node_ids = set(self._inference_nodes)
+            for ctr in conn.execute("SELECT payload FROM compare_tasks").fetchall():
+                ls = json_loads(ctr["payload"], {}).get("load_status") or {}
+                if isinstance(ls, str):
+                    try:
+                        ls = json.loads(ls)
+                    except (json.JSONDecodeError, TypeError):
+                        ls = {}
+                for m in ls.get("loaded_models") or []:
+                    if m.get("status") in {"ready", "running"} and m.get("node_id"):
+                        inference_node_ids.add(m["node_id"])
             items = []
             for row in rows:
                 task = next(
@@ -2911,11 +3106,21 @@ class PlatformStore:
                     ),
                     None,
                 )
-                busy = task is not None and task.get("status") == "running"
-                reserved = task is not None and task.get("status") in {"syncing", "queued"}
+                eval_task = next(
+                    (
+                        t
+                        for t in eval_running
+                        if t.get("compute_node_id") == row["node_id"]
+                        and row["gpu_index"] == (int(t["gpu_id"]) if t.get("gpu_id") is not None else -1)
+                    ),
+                    None,
+                )
+                busy = (task is not None and task.get("status") == "running") or eval_task is not None
+                reserved = (task is not None and task.get("status") in {"syncing", "queued"}) or (
+                    eval_task is not None and eval_task.get("status") in {"syncing", "queued"}
+                )
                 # Also mark GPU as busy if an inference model is loaded on this node
-                inference_busy = self.is_inference_loaded(row["node_id"])
-                if inference_busy and not busy:
+                if row["node_id"] in inference_node_ids and not busy:
                     busy = True
                     reserved = False
                 memory_used = round(row["memory_total_gb"] * (0.72 if busy else 0.18 if reserved else 0.04), 1)
@@ -2947,6 +3152,16 @@ class PlatformStore:
                             }
                         ]
                         if task
+                        else [
+                            {
+                                "pid": int(eval_task.get("process_id") or 0),
+                                "name": "eval_runner",
+                                "memory_used_gb": memory_used,
+                                "task_name": eval_task.get("eval_task_name") or eval_task.get("name") or "评测任务",
+                                "user": "admin",
+                            }
+                        ]
+                        if eval_task
                         else [],
                     }
                 )

backend/app/db/sql/001_platform_runtime.sql

@@ -33,7 +33,10 @@ CREATE TABLE IF NOT EXISTS trained_models (
     create_time TEXT NOT NULL,
     merged INTEGER NOT NULL DEFAULT 0,
     merging INTEGER NOT NULL DEFAULT 0,
-    merged_path TEXT
+    merged_path TEXT,
+    artifact_dir TEXT,
+    compute_node_id TEXT,
+    compute_node_name TEXT
 );
 
 CREATE TABLE IF NOT EXISTS model_lineage (

backend/app/modules/compute_gateway/client.py

@@ -33,6 +33,16 @@ def _unwrap_dict(payload: Any) -> dict[str, Any]:
     return payload if isinstance(payload, dict) else {}
 
 
+# Inference calls are intentionally short-timeout:
+#   - load dispatch only confirms the compute node accepted the request
+#     (the actual model load now runs asynchronously on the node).
+#   - status/unload must never block the platform for long when a node is
+#     unreachable but still marked online.
+INFERENCE_LOAD_TIMEOUT = httpx.Timeout(30, connect=10)
+INFERENCE_STATUS_TIMEOUT = httpx.Timeout(30, connect=5)
+INFERENCE_UNLOAD_TIMEOUT = httpx.Timeout(30, connect=5)
+
+
 class ComputeNodeClient:
     """Application-side client for one compute node.
 
@@ -182,10 +192,16 @@ class ComputeNodeClient:
             response.raise_for_status()
             return _unwrap_dict(response.json())
 
-    async def _request(self, method: str, path: str, json_data: dict[str, Any] | None = None) -> dict[str, Any]:
+    async def _request(
+        self,
+        method: str,
+        path: str,
+        json_data: dict[str, Any] | None = None,
+        timeout: float | None = None,
+    ) -> dict[str, Any]:
         """Generic request method for compute API endpoints."""
         url = _join_url(self.api_base_url, f"{self.route_prefix}{path}")
-        async with httpx.AsyncClient(timeout=300, headers=self.headers()) as client:
+        async with httpx.AsyncClient(timeout=timeout or 300, headers=self.headers()) as client:
             if method.upper() == "GET":
                 response = await client.get(url)
             else:
@@ -193,6 +209,19 @@ class ComputeNodeClient:
             response.raise_for_status()
             return _unwrap_dict(response.json())
 
+    # ── Inference helpers (short timeouts — see module constants) ──────────
+
+    async def inference_load(self, payload: dict[str, Any]) -> dict[str, Any]:
+        """Dispatch a model load. Returns as soon as the node accepts the
+        request; the node now loads asynchronously (status goes 'loading')."""
+        return await self._request("POST", "/inference/load", json_data=payload, timeout=INFERENCE_LOAD_TIMEOUT)
+
+    async def inference_status(self) -> dict[str, Any]:
+        return await self._request("GET", "/inference/status", timeout=INFERENCE_STATUS_TIMEOUT)
+
+    async def inference_unload(self) -> dict[str, Any]:
+        return await self._request("POST", "/inference/unload", json_data={}, timeout=INFERENCE_UNLOAD_TIMEOUT)
+
     async def upload_file(
         self,
         filename: str,

backend/app/modules/compute_gateway/sync.py

@@ -1,15 +1,101 @@
 from __future__ import annotations
 
+import json
+import time
 from typing import Any
 
 from app.db.platform_store import get_platform_store
 from app.modules.compute_gateway.client import ComputeNodeClient
 
+# starting 状态允许的最大轮询次数（约 40 * 3s ≈ 2 分钟），超过即判定节点不可达
+MAX_STARTING_ATTEMPTS = 40
+
 
 def _node_for_task(task: dict[str, Any]) -> dict[str, Any] | None:
     return next((node for node in get_platform_store().compute_nodes() if node["id"] == task.get("compute_node_id")), None)
 
 
+def _parse_inference_load_status(task: dict[str, Any]) -> tuple[list[dict[str, Any]], dict[str, Any]]:
+    load_status = task.get("load_status") or {}
+    if isinstance(load_status, str):
+        try:
+            load_status = json.loads(load_status)
+        except (json.JSONDecodeError, TypeError):
+            load_status = {}
+    return load_status.get("loaded_models") or [], load_status
+
+
+async def reconcile_inference_loads(store: Any) -> list[dict[str, Any]]:
+    """推进处于 starting 状态的推理加载。
+
+    模型加载已改为异步派发：/model-compare/{id}/load 立即返回，这里在每次
+    轮询时查询对应计算节点的 /inference/status，把任务从 starting 推进到
+    ready/error。使用短超时，单节点不可达不会阻塞整轮轮询。
+    """
+    reconciled: list[dict[str, Any]] = []
+    now = time.time()
+    for task in store.compare_tasks():
+        items, _ = _parse_inference_load_status(task)
+        if not any(item.get("status") == "starting" for item in items):
+            continue
+        # dirty 只要处理过任一 starting 项就置位：load_attempts / last_polled_at
+        # 必须落库，否则节点不可达时计数不会累积，封顶逻辑永远触发不了
+        dirty = False
+        for item in items:
+            if item.get("status") != "starting":
+                continue
+            # 节流：同一 item 每 3s 只查询一次
+            if now - float(item.get("last_polled_at") or 0) < 3:
+                continue
+            item["last_polled_at"] = now
+            item["load_attempts"] = int(item.get("load_attempts") or 0) + 1
+            dirty = True
+            node = next((n for n in store.compute_nodes() if n["id"] == item.get("node_id")), None)
+            if not node:
+                item["status"] = "error"
+                item["error"] = "compute node deleted"
+                store.mark_inference_unloaded(item.get("node_id") or "")
+                continue
+            if not node.get("enabled") or node.get("scheduler_status") != "online":
+                item["status"] = "error"
+                item["error"] = "compute node offline"
+                store.mark_inference_unloaded(node["id"])
+                continue
+            try:
+                status = await ComputeNodeClient(node["api_base_url"]).inference_status()
+            except Exception as exc:  # noqa: BLE001 - node unreachable; keep retrying until cap
+                if int(item.get("load_attempts") or 0) >= MAX_STARTING_ATTEMPTS:
+                    item["status"] = "error"
+                    item["error"] = f"compute node unreachable: {exc}"
+                    store.mark_inference_unloaded(node["id"])
+                continue
+            node_status = status.get("status")
+            if node_status == "ready":
+                item["status"] = "ready"
+                item.pop("error", None)
+                store.mark_inference_loaded(node["id"])
+            elif node_status == "error":
+                item["status"] = "error"
+                item["error"] = status.get("error") or "model load failed on compute node"
+                store.mark_inference_unloaded(node["id"])
+            elif node_status == "idle":
+                # 节点重启导致已加载模型丢失
+                item["status"] = "error"
+                item["error"] = "model disappeared from compute node (node may have restarted)"
+                store.mark_inference_unloaded(node["id"])
+            # node_status == "loading" -> 保持 starting，下轮再查
+        if dirty:
+            if any(i.get("status") in {"ready", "running"} for i in items):
+                new_status = "loaded"
+            elif any(i.get("status") == "starting" for i in items):
+                new_status = "starting"  # 仍在加载中，保持 starting
+            else:
+                new_status = "failed"
+            store.update_compare_task(task["id"], {"status": new_status, "load_status": {"loaded_models": items}})
+            reconciled.append({"task_id": task["id"], "status": new_status})
+    return reconciled
+
+
 async def fetch_eval_result_content(client: ComputeNodeClient, node: dict[str, Any], job: dict[str, Any]) -> dict[str, Any] | None:
     output_dir = job.get("output_dir")
     if not output_dir:
@@ -88,12 +174,18 @@ async def poll_compute_jobs_once() -> dict[str, Any]:
                 except Exception:
                     pass
             store.apply_eval_job_result(eval_task["id"], job, result_content)
-            # If job completed, un-mark inference loaded
-            if job.get("status") in {"completed", "failed", "stopped"}:
-                store.mark_inference_unloaded(node["id"])
+            # 评测 GPU 占用由 eval_tasks 状态派生，无需维护推理内存标记
             eval_synced += 1
         except Exception as exc:  # noqa: BLE001
             failed.append({"eval_task_id": eval_task["id"], "error": str(exc)})
 
+    # ── Inference load reconciliation ─────────────────────────────────────
+    try:
+        inference_reconciled = await reconcile_inference_loads(store)
+    except Exception as exc:  # noqa: BLE001 - keep polling alive
+        failed.append({"inference_reconcile": str(exc)})
+        inference_reconciled = []
+
     return {"synced": len(synced) + len(standalone_synced) + eval_synced, "failed": failed,
-            "items": synced, "standalone": standalone_synced, "eval_synced": eval_synced}
+            "items": synced, "standalone": standalone_synced, "eval_synced": eval_synced,
+            "inference_reconciled": inference_reconciled}

backend/requirements.txt

@@ -19,3 +19,7 @@ llama-index-core==0.14.23
 llama-index-embeddings-huggingface==0.6.1
 docling==2.115.0
 tiktoken>=0.7.0
+
+# 测试与代码检查
+pytest>=8.2.0
+ruff>=0.5.0

backend/tests/test_compare_inference_async.py

@@ -0,0 +1,276 @@
+"""
+模型推理异步加载改造的单元测试。
+
+覆盖：
+- model_compare_load：异步派发，立即返回 starting + 节点信息（不等待加载完成）
+- model_compare_delete：先删记录，卸载失败也不阻塞删除
+- reconcile_inference_loads：starting -> ready/error/idle/不可达的状态迁移与封顶
+- _unload_from_compute_node：任务感知，只命中记录中的节点
+"""
+
+from __future__ import annotations
+
+import asyncio
+from types import SimpleNamespace
+from typing import Any
+
+from app.api.v1.endpoints.platform import model_compare_delete, model_compare_load
+import app.api.v1.endpoints.platform as platform
+from app.modules.compute_gateway.client import ComputeNodeClient
+from app.modules.compute_gateway.sync import MAX_STARTING_ATTEMPTS, reconcile_inference_loads
+
+
+class FakeInferenceStore:
+    """内存 store，仅实现推理加载/对账用到的接口。"""
+
+    def __init__(self, tasks: list[dict[str, Any]] | None = None, nodes: list[dict[str, Any]] | None = None) -> None:
+        self._tasks: dict[str, dict[str, Any]] = {t["id"]: dict(t) for t in (tasks or [])}
+        self._nodes = nodes or []
+        self._inference_nodes: set[str] = set()
+
+    def compare_task(self, task_id: str) -> dict[str, Any]:
+        if task_id not in self._tasks:
+            raise KeyError(task_id)
+        return dict(self._tasks[task_id])
+
+    def compare_tasks(self) -> list[dict[str, Any]]:
+        return [dict(t) for t in self._tasks.values()]
+
+    def update_compare_task(self, task_id: str, payload: dict[str, Any]) -> dict[str, Any]:
+        current = self._tasks[task_id]
+        merged = {**current, **payload, "id": task_id}
+        self._tasks[task_id] = merged
+        return dict(merged)
+
+    def delete_compare_task(self, task_id: str) -> None:
+        self._tasks.pop(task_id, None)
+
+    def compute_nodes(self) -> list[dict[str, Any]]:
+        return [dict(n) for n in self._nodes]
+
+    def model(self, model_id: str) -> dict[str, Any]:
+        raise KeyError(model_id)
+
+    def trained_models(self) -> list[dict[str, Any]]:
+        return []
+
+    def mark_inference_loaded(self, node_id: str) -> None:
+        self._inference_nodes.add(node_id)
+
+    def mark_inference_unloaded(self, node_id: str) -> None:
+        self._inference_nodes.discard(node_id)
+
+    def is_inference_loaded(self, node_id: str) -> bool:
+        return node_id in self._inference_nodes
+
+
+def _node(node_id: str, code: str = "") -> dict[str, Any]:
+    return {
+        "id": node_id,
+        "code": code or node_id,
+        "name": code or node_id,
+        "api_base_url": f"http://{code or node_id}:19100",
+        "enabled": True,
+        "scheduler_status": "online",
+    }
+
+
+def _task(task_id: str, *, node_id: str | None = None, load_status: dict[str, Any] | None = None) -> dict[str, Any]:
+    return {
+        "id": task_id,
+        "name": f"task-{task_id}",
+        "status": "pending",
+        "models": [
+            {"model_id": "m_1", "model_name": "qwen", "model_path": "/models/qwen", "node_id": node_id}
+        ],
+        "load_status": load_status or {"loaded_models": []},
+    }
+
+
+async def _fake_inference_load(self, payload: dict[str, Any]) -> dict[str, Any]:
+    return {"loaded": False, "status": "loading", "request_id": "req-1"}
+
+
+async def _fake_inference_unload(self) -> dict[str, Any]:
+    return {"unloaded": True, "status": "idle"}
+
+
+def _patch_store(monkeypatch, store: FakeInferenceStore) -> None:
+    monkeypatch.setattr(platform, "get_platform_store", lambda: store)
+    monkeypatch.setattr(platform, "get_settings", lambda: SimpleNamespace(compute_mode="real"))
+
+
+def test_select_eval_node_prefers_model_node(monkeypatch) -> None:
+    from app.api.v1.endpoints.platform import _select_eval_node
+
+    store = FakeInferenceStore(nodes=[_node("n1"), _node("n2")])
+    # 指定模型所在节点时优先返回该节点
+    assert _select_eval_node(store, "n2")["id"] == "n2"
+    # 无指定节点时回退到第一个在线节点
+    assert _select_eval_node(store, None)["id"] == "n1"
+
+
+def test_select_eval_node_returns_none_when_model_node_offline(monkeypatch) -> None:
+    from app.api.v1.endpoints.platform import _select_eval_node
+
+    nodes = [_node("n1"), _node("n2")]
+    nodes[1]["enabled"] = False
+    store = FakeInferenceStore(nodes=nodes)
+    # 模型所在节点不可用 → 明确失败，不派发到其它节点
+    assert _select_eval_node(store, "n2") is None
+    # 无指定节点时仍回退第一个在线节点
+    assert _select_eval_node(store, None)["id"] == "n1"
+
+
+def test_model_compare_load_dispatches_and_returns_starting(monkeypatch) -> None:
+    store = FakeInferenceStore(tasks=[_task("t1", node_id="n1")], nodes=[_node("n1")])
+    _patch_store(monkeypatch, store)
+    monkeypatch.setattr(ComputeNodeClient, "inference_load", _fake_inference_load)
+
+    result = asyncio.run(model_compare_load("t1"))
+    assert result["code"] == 0
+    updated = result["data"]
+    assert updated["status"] == "starting"
+    items = updated["load_status"]["loaded_models"]
+    assert items[0]["status"] == "starting"
+    assert items[0]["node_id"] == "n1"
+    assert "n1" in store._inference_nodes
+
+
+def test_model_compare_load_marks_error_when_all_nodes_fail(monkeypatch) -> None:
+    store = FakeInferenceStore(tasks=[_task("t1", node_id="n1")], nodes=[_node("n1")])
+    _patch_store(monkeypatch, store)
+
+    async def _raise(self, payload: dict[str, Any]) -> dict[str, Any]:
+        raise RuntimeError("conn refused")
+
+    monkeypatch.setattr(ComputeNodeClient, "inference_load", _raise)
+
+    result = asyncio.run(model_compare_load("t1"))
+    updated = result["data"]
+    assert updated["status"] == "failed"
+    assert updated["load_status"]["loaded_models"][0]["status"] == "error"
+    assert "conn refused" in updated["load_status"]["loaded_models"][0]["error"]
+
+
+def test_model_compare_delete_removes_record_even_if_unload_raises(monkeypatch) -> None:
+    task = _task(
+        "t1",
+        node_id="n1",
+        load_status={"loaded_models": [{"model_id": "m_1", "status": "ready", "node_id": "n1"}]},
+    )
+    store = FakeInferenceStore(tasks=[task], nodes=[_node("n1")])
+    _patch_store(monkeypatch, store)
+
+    async def _raise(self) -> dict[str, Any]:
+        raise RuntimeError("unload boom")
+
+    monkeypatch.setattr(ComputeNodeClient, "inference_unload", _raise)
+
+    result = asyncio.run(model_compare_delete("t1"))
+    assert result["data"] == {"deleted": "t1"}
+    assert "t1" not in store._tasks
+    # finally 中仍清掉了节点标记
+    assert "n1" not in store._inference_nodes
+
+
+def test_unload_from_compute_node_only_hits_recorded_node(monkeypatch) -> None:
+    task = _task(
+        "t1",
+        load_status={"loaded_models": [{"model_id": "m_1", "status": "ready", "node_id": "n1"}]},
+    )
+    store = FakeInferenceStore(tasks=[task], nodes=[_node("n1"), _node("n2")])
+    _patch_store(monkeypatch, store)
+    monkeypatch.setattr(ComputeNodeClient, "inference_unload", _fake_inference_unload)
+
+    from app.api.v1.endpoints.platform import _unload_from_compute_node
+
+    result = asyncio.run(_unload_from_compute_node(store, task=task))
+    assert result["unloaded"] is True
+    # 只命中任务记录中的节点 n1，n2 未被卸载
+    assert [r["node_id"] for r in result["nodes"]] == ["n1"]
+    assert "n1" not in store._inference_nodes
+
+
+async def _status_ready(self) -> dict[str, Any]:
+    return {"loaded": True, "status": "ready", "model_name": "qwen"}
+
+
+def test_reconcile_transitions_starting_to_ready(monkeypatch) -> None:
+    task = _task(
+        "t1",
+        node_id="n1",
+        load_status={"loaded_models": [{"model_id": "m_1", "status": "starting", "node_id": "n1"}]},
+    )
+    store = FakeInferenceStore(tasks=[task], nodes=[_node("n1")])
+    monkeypatch.setattr(ComputeNodeClient, "inference_status", _status_ready)
+
+    reconciled = asyncio.run(reconcile_inference_loads(store))
+    assert reconciled == [{"task_id": "t1", "status": "loaded"}]
+    updated = store._tasks["t1"]
+    assert updated["status"] == "loaded"
+    assert updated["load_status"]["loaded_models"][0]["status"] == "ready"
+    assert "n1" in store._inference_nodes
+
+
+def test_reconcile_transitions_to_error_and_failed(monkeypatch) -> None:
+    async def _status_error(self) -> dict[str, Any]:
+        return {"loaded": False, "status": "error", "error": "CUDA out of memory"}
+
+    task = _task(
+        "t1",
+        node_id="n1",
+        load_status={"loaded_models": [{"model_id": "m_1", "status": "starting", "node_id": "n1"}]},
+    )
+    store = FakeInferenceStore(tasks=[task], nodes=[_node("n1")])
+    monkeypatch.setattr(ComputeNodeClient, "inference_status", _status_error)
+
+    reconciled = asyncio.run(reconcile_inference_loads(store))
+    assert reconciled == [{"task_id": "t1", "status": "failed"}]
+    item = store._tasks["t1"]["load_status"]["loaded_models"][0]
+    assert item["status"] == "error"
+    assert "CUDA out of memory" in item["error"]
+    assert "n1" not in store._inference_nodes
+
+
+def test_reconcile_idle_marks_model_disappeared(monkeypatch) -> None:
+    async def _status_idle(self) -> dict[str, Any]:
+        return {"loaded": False, "status": "idle"}
+
+    task = _task(
+        "t1",
+        node_id="n1",
+        load_status={"loaded_models": [{"model_id": "m_1", "status": "starting", "node_id": "n1"}]},
+    )
+    store = FakeInferenceStore(tasks=[task], nodes=[_node("n1")])
+    monkeypatch.setattr(ComputeNodeClient, "inference_status", _status_idle)
+
+    asyncio.run(reconcile_inference_loads(store))
+    item = store._tasks["t1"]["load_status"]["loaded_models"][0]
+    assert item["status"] == "error"
+    assert "disappeared" in item["error"]
+    assert store._tasks["t1"]["status"] == "failed"
+
+
+def test_reconcile_unreachable_node_flips_to_error_after_cap(monkeypatch) -> None:
+    async def _raise(self) -> dict[str, Any]:
+        raise RuntimeError("conn refused")
+
+    task = _task(
+        "t1",
+        node_id="n1",
+        load_status={"loaded_models": [{"model_id": "m_1", "status": "starting", "node_id": "n1"}]},
+    )
+    store = FakeInferenceStore(tasks=[task], nodes=[_node("n1")])
+    monkeypatch.setattr(ComputeNodeClient, "inference_status", _raise)
+
+    # 每次轮询前重置节流时间戳，逐次推进 load_attempts 到封顶
+    for _ in range(MAX_STARTING_ATTEMPTS):
+        item = store._tasks["t1"]["load_status"]["loaded_models"][0]
+        item["last_polled_at"] = 0
+        asyncio.run(reconcile_inference_loads(store))
+
+    item = store._tasks["t1"]["load_status"]["loaded_models"][0]
+    assert item["status"] == "error"
+    assert "unreachable" in item["error"]
+    assert store._tasks["t1"]["status"] == "failed"

compute/api/main.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import asyncio
 import json
 import os
 import math
@@ -717,7 +718,9 @@ def create_app() -> FastAPI:
     @app.post(f"{route_prefix}/inference/unload")
     async def inference_unload() -> dict[str, Any]:
         """Unload the currently loaded model and free GPU memory."""
-        return get_inference_session().unload()
+        # Teardown (gc.collect + cuda.empty_cache) can take a while; run it off
+        # the event loop so /health and /inference/status stay responsive.
+        return await asyncio.to_thread(get_inference_session().unload)
 
     @app.get(f"{route_prefix}/inference/status")
     async def inference_status() -> dict[str, Any]:
@@ -737,7 +740,10 @@ def create_app() -> FastAPI:
         messages = payload.get("messages") or []
         if not messages:
             raise HTTPException(status_code=400, detail="messages is required")
-        result = get_inference_session().chat(
+        # Generation is long-running; run it in a thread so the event loop keeps
+        # serving /inference/status and /health during inference.
+        result = await asyncio.to_thread(
+            get_inference_session().chat,
             messages=messages,
             temperature=float(payload.get("temperature", 0.95)),
             top_p=float(payload.get("top_p", 0.7)),

compute/engines/llama_factory/eval_runner.py

@@ -183,6 +183,9 @@ def _judge_sample(
     api_url = (config.get("api_url") or "").strip().rstrip("/")
     api_key = (config.get("api_key") or "").strip()
     eval_model = (config.get("eval_model") or "").strip()
+    # 优先使用模型记录里配置的真实 API 模型名（如 deepseek-chat），
+    # 否则回退到平台内部模型名
+    api_model = (config.get("api_model") or "").strip() or eval_model
     eval_prompt = (config.get("eval_prompt") or "").strip()
     score_min = float(config.get("score_min", 0))
     score_max = float(config.get("score_max", 5))
@@ -208,7 +211,7 @@ def _judge_sample(
         import urllib.error
 
         body = json.dumps({
-            "model": eval_model,
+            "model": api_model,
             "messages": [
                 {"role": "system", "content": system_msg},
                 {"role": "user", "content": user_msg},
@@ -308,13 +311,15 @@ def run_eval(config: dict[str, Any]) -> dict[str, Any]:
     print(f"[eval] loading model: {model_path}")
     from compute.engines.llama_factory.inference import InferenceSession
     session = InferenceSession()
-    load_result = session.load(
+    session.load(
         model_name_or_path=model_path,
         adapter_name_or_path=adapter_path,
         template=template,
         infer_backend=config.get("infer_backend", "huggingface"),
         infer_dtype=config.get("infer_dtype", "auto"),
     )
+    # load() 为异步加载（立即返回 loading），必须等待后台线程完成后再进行推理
+    load_result = session.wait_until_loaded(timeout=float(config.get("load_timeout", 1800)))
     if not load_result.get("loaded"):
         raise RuntimeError(f"model load failed: {load_result.get('error', 'unknown')}")
     print(f"[eval] model loaded OK")

compute/engines/llama_factory/inference.py

@@ -2,114 +2,233 @@ from __future__ import annotations
 
 import threading
 import time
-from typing import Any
+import uuid
+from typing import Any, Iterator
 
 
 class InferenceSession:
-    """Manages a loaded model for inference with LLaMA-Factory ChatModel."""
+    """Manages a loaded model for inference with LLaMA-Factory ChatModel.
+
+    Model loading is asynchronous: ``load()`` spawns a background daemon thread
+    and returns immediately with ``status == "loading"``. ``info()`` (served by
+    ``/inference/status``) is always responsive, so the platform backend can
+    poll loading progress without being blocked by a minutes-long model load —
+    which previously froze the whole compute node event loop.
+
+    State machine: idle -> loading -> ready | error, ready -> idle (unload),
+    loading -> idle (cancelled). Long operations (ChatModel build, teardown,
+    generation) never run while holding ``_state_lock``; they either run in the
+    worker thread or under ``_chat_lock`` only.
+    """
 
     def __init__(self) -> None:
+        self._state_lock = threading.Lock()  # brief state transitions only
+        self._chat_lock = threading.Lock()  # serialize chat/teardown
+        self._status: str = "idle"
+        self._error: str = ""
+        self._request_id: str = ""
+        self._load_args: dict[str, Any] = {}
+        self._teardown_old = False  # load-while-ready: unload old before loading new
+        self._cancel_requested = False  # unload-while-loading: tear down after load finishes
+        self._load_thread: threading.Thread | None = None
         self._model: Any = None
         self._tokenizer: Any = None
         self._generating_args: dict[str, Any] = {}
         self._model_name: str = ""
         self._adapter_path: str = ""
-        self._lock = threading.Lock()
         self._loaded_at: float = 0.0
-        self._status: str = "idle"
 
     @property
     def status(self) -> str:
+        with self._state_lock:
             return self._status
 
-    @property
-    def model_name(self) -> str:
-        return self._model_name
-
-    @property
-    def adapter_path(self) -> str:
-        return self._adapter_path
-
-    @property
-    def loaded_at(self) -> float:
-        return self._loaded_at
-
     def info(self) -> dict[str, Any]:
+        with self._state_lock:
             return {
                 "loaded": self._status == "ready",
                 "status": self._status,
                 "model_name": self._model_name,
                 "adapter_path": self._adapter_path,
                 "loaded_at": self._loaded_at,
+                "request_id": self._request_id,
+                "error": self._error,
             }
 
-    def load(self, model_name_or_path, adapter_name_or_path="", template="qwen", infer_backend="huggingface", infer_dtype="auto", **kwargs):
-        with self._lock:
+    def wait_until_loaded(self, timeout: float | None = None) -> dict[str, Any]:
+        """Wait for an in-flight async load to finish and return its outcome.
+
+        供同步消费方（如 eval_runner 子进程）使用：``load()`` 立即返回 loading 后，
+        调用本方法等待后台加载线程完成，拿到最终的 loaded/error 结果。
+        若在 timeout 秒内仍未加载完成，返回 ``status == "loading"`` 并附上超时提示。
+        """
+        with self._state_lock:
+            thread = self._load_thread
+        if thread is not None and thread.is_alive():
+            thread.join(timeout=timeout)
+        with self._state_lock:
+            loaded = self._status == "ready"
+            status = self._status
+            error = self._error
+        if not loaded and status == "loading":
+            error = error or f"model load timed out after {timeout or 'N/A'}s"
+        return {
+            "loaded": loaded,
+            "status": status,
+            "model_name": self._model_name,
+            "adapter_path": self._adapter_path,
+            "error": error,
+        }
+
+    def load(
+        self,
+        model_name_or_path,
+        adapter_name_or_path="",
+        template="qwen",
+        infer_backend="huggingface",
+        infer_dtype="auto",
+        **kwargs,
+    ) -> dict[str, Any]:
+        with self._state_lock:
             if self._status == "loading":
-                return {"loaded": False, "error": "model is already loading"}
-            if self._status == "ready":
-                self.unload()
+                # A model is already loading — dedupe, reuse the same request id.
+                return {"loaded": False, "status": "loading", "request_id": self._request_id}
+            self._teardown_old = self._status == "ready"
             self._status = "loading"
+            self._error = ""
+            self._request_id = uuid.uuid4().hex[:12]
+            self._cancel_requested = False
+            self._load_args = {
+                "model_name_or_path": model_name_or_path,
+                "template": template,
+                "infer_backend": infer_backend,
+                "infer_dtype": infer_dtype,
+            }
+            if adapter_name_or_path:
+                self._load_args["adapter_name_or_path"] = adapter_name_or_path
+            self._load_args.update(kwargs)
             self._model_name = model_name_or_path
             self._adapter_path = adapter_name_or_path
+            self._load_thread = threading.Thread(target=self._load_worker, daemon=True)
+            self._load_thread.start()
+            return {"loaded": False, "status": "loading", "request_id": self._request_id}
+
+    def _load_worker(self) -> None:
+        """Build the ChatModel off the state lock so info() never blocks."""
+        model = None
+        tokenizer = None
+        generating_args: dict[str, Any] = {}
+        error = ""
         try:
+            if self._teardown_old:
+                self._release_model()
             from llamafactory.chat import ChatModel
             from llamafactory.hparams import get_infer_args
-                args = {"model_name_or_path": model_name_or_path, "template": template, "infer_backend": infer_backend, "infer_dtype": infer_dtype}
-                if adapter_name_or_path:
-                    args["adapter_name_or_path"] = adapter_name_or_path
-                args.update(kwargs)
-                infer_result = get_infer_args(args)
-                # ChatModel internally re-parses the args dict via get_infer_args,
-                # so pass the original args (not the parsed dataclass objects).
-                self._model = ChatModel(args)
-                self._tokenizer = getattr(self._model, 'tokenizer', None) or self._model.engine.tokenizer
-                # Extract generating_args (last element) for later use in chat()
-                generating_args = infer_result[-1]
-                if hasattr(generating_args, '__dataclass_fields__'):
-                    self._generating_args = {k: v for k, v in vars(generating_args).items()
-                                              if not k.startswith('_')}
-                else:
-                    self._generating_args = dict(generating_args)
-                self._loaded_at = time.time()
-                self._status = "ready"
-                return {"loaded": True, "status": "ready"}
-            except Exception as exc:
-                self._status = "error"
-                self._model = None
-                return {"loaded": False, "status": "error", "error": str(exc)}
 
-    def unload(self):
-        with self._lock:
-            if self._model is not None:
-                try:
-                    del self._model
-                except Exception:
-                    pass
+            args = dict(self._load_args)
+            infer_result = get_infer_args(args)
+            model = ChatModel(args)
+            tokenizer = getattr(model, "tokenizer", None) or model.engine.tokenizer
+            generating_args = infer_result[-1]
+            if hasattr(generating_args, "__dataclass_fields__"):
+                generating_args = {
+                    k: v for k, v in vars(generating_args).items() if not k.startswith("_")
+                }
+            else:
+                generating_args = dict(generating_args)
+        except Exception as exc:  # noqa: BLE001 - surface load failure via status
+            error = str(exc)
+        with self._state_lock:
+            if error:
                 self._model = None
                 self._tokenizer = None
+                self._status = "error"
+                self._error = error
+                return
+            if self._cancel_requested:
+                # Unload was requested while loading — drop the fresh model.
+                model = None
+                tokenizer = None
+                self._model = None
+                self._tokenizer = None
+                self._status = "idle"
+                return
+            self._model = model
+            self._tokenizer = tokenizer
+            self._generating_args = generating_args
+            self._loaded_at = time.time()
+            self._status = "ready"
+
+    def _release_model(self) -> None:
+        with self._chat_lock:
+            with self._state_lock:
+                self._status = "unloading"
+                model = self._model
+                self._model = None
+                self._tokenizer = None
+            if model is not None:
+                try:
+                    del model
+                except Exception:  # noqa: BLE001 - best-effort teardown
+                    pass
                 # 强制释放 PyTorch CUDA 缓存，真正归还 GPU 显存
                 try:
                     import gc
+
                     gc.collect()
                     import torch
+
                     if torch.cuda.is_available():
                         torch.cuda.empty_cache()
                         torch.cuda.synchronize()
-                except Exception:
+                except Exception:  # noqa: BLE001 - teardown must not raise
                     pass
+            with self._state_lock:
                 self._status = "idle"
                 self._model_name = ""
                 self._adapter_path = ""
                 self._loaded_at = 0.0
-            return {"unloaded": True}
+                self._error = ""
 
-    def chat(self, messages, temperature=0.95, top_p=0.7, max_new_tokens=1024, do_sample=True, **kwargs):
-        with self._lock:
+    def unload(self) -> dict[str, Any]:
+        with self._state_lock:
+            if self._status == "loading":
+                # Ask the worker to tear down right after the load finishes.
+                self._cancel_requested = True
+                return {"unloaded": False, "status": "cancelling", "request_id": self._request_id}
+            was_ready = self._status == "ready"
+        if was_ready:
+            self._release_model()
+        else:
+            with self._state_lock:
+                self._model = None
+                self._tokenizer = None
+                self._status = "idle"
+                self._model_name = ""
+                self._adapter_path = ""
+                self._loaded_at = 0.0
+                self._error = ""
+        return {"unloaded": True, "status": "idle"}
+
+    def chat(self, messages, temperature=0.95, top_p=0.7, max_new_tokens=1024, do_sample=True, **kwargs) -> dict[str, Any]:
+        with self._chat_lock:
+            with self._state_lock:
+                if self._status == "loading":
+                    return {
+                        "error": f"model is still loading (request_id={self._request_id}); please retry",
+                        "response": "",
+                    }
+                if self._status == "error":
+                    return {"error": f"model load failed: {self._error}", "response": ""}
                 if self._status != "ready" or self._model is None:
                     return {"error": "model not loaded", "response": ""}
             try:
-                generate_kwargs = {"temperature": temperature, "top_p": top_p, "max_new_tokens": max_new_tokens, "do_sample": do_sample}
+                generate_kwargs = {
+                    "temperature": temperature,
+                    "top_p": top_p,
+                    "max_new_tokens": max_new_tokens,
+                    "do_sample": do_sample,
+                }
                 generate_kwargs.update(kwargs)
                 system = next((m["content"] for m in messages if m["role"] == "system"), None)
                 user_messages = [m for m in messages if m["role"] != "system"]
@@ -118,11 +237,18 @@ class InferenceSession:
                     responses.append(response)
                 full_response = "".join(str(r) for r in responses)
                 return {"response": full_response}
-            except Exception as exc:
+            except Exception as exc:  # noqa: BLE001 - return generation error to caller
                 return {"error": str(exc), "response": ""}
 
-    def chat_stream(self, messages, **kwargs):
-        with self._lock:
+    def chat_stream(self, messages, **kwargs) -> Iterator[str]:
+        with self._chat_lock:
+            with self._state_lock:
+                if self._status == "loading":
+                    yield 'data: {"error": "model is still loading; please retry"}\n\n'
+                    return
+                if self._status == "error":
+                    yield 'data: {"error": "model load failed: ' + str(self._error) + '"}\n\n'
+                    return
                 if self._status != "ready" or self._model is None:
                     yield 'data: {"error": "model not loaded"}\n\n'
                     return
@@ -132,13 +258,14 @@ class InferenceSession:
                 user_messages = [m for m in messages if m["role"] != "system"]
                 for new_text in self._model.stream_chat(user_messages, system=system, **generate_kwargs):
                     yield new_text
-            except Exception as exc:
+            except Exception as exc:  # noqa: BLE001 - stream error as SSE event
                 yield 'data: {"error": "' + str(exc) + '"}\n\n'
 
 
 _inference_session = None
 
-def get_inference_session():
+
+def get_inference_session() -> InferenceSession:
     global _inference_session
     if _inference_session is None:
         _inference_session = InferenceSession()

compute/tests/test_inference_session.py

@@ -0,0 +1,146 @@
+from __future__ import annotations
+
+import sys
+import time
+import types
+from typing import Any
+
+import pytest
+
+from compute.engines.llama_factory.inference import InferenceSession
+
+# 模拟模型加载耗时，用于验证 load() 立即返回、info() 不阻塞
+LOAD_DELAY = 0.2
+
+
+class FakeChatModel:
+    def __init__(self, args: dict[str, Any]) -> None:
+        time.sleep(LOAD_DELAY)
+        self.tokenizer = object()
+        self.engine = types.SimpleNamespace(tokenizer=object())
+        self._output = "hello from model"
+
+    def stream_chat(self, *args, **kwargs):
+        for _ in range(1):
+            yield self._output
+
+
+class FailingChatModel:
+    def __init__(self, args: dict[str, Any]) -> None:
+        time.sleep(LOAD_DELAY)
+        raise RuntimeError("boom: fake load failure")
+
+
+def _get_infer_args(args: dict[str, Any]) -> list[Any]:
+    # 最后一个元素为 generating_args，worker 会转成 dict
+    return [None, None, {"temperature": 0.7}]
+
+
+def _install_llamafactory(monkeypatch, chat_model: type) -> None:
+    llmf = types.ModuleType("llamafactory")
+    chat_mod = types.ModuleType("llamafactory.chat")
+    hparams_mod = types.ModuleType("llamafactory.hparams")
+    chat_mod.ChatModel = chat_model
+    hparams_mod.get_infer_args = _get_infer_args
+    llmf.chat = chat_mod
+    llmf.hparams = hparams_mod
+    monkeypatch.setitem(sys.modules, "llamafactory", llmf)
+    monkeypatch.setitem(sys.modules, "llamafactory.chat", chat_mod)
+    monkeypatch.setitem(sys.modules, "llamafactory.hparams", hparams_mod)
+
+
+@pytest.fixture
+def stub_llamafactory(monkeypatch) -> None:
+    _install_llamafactory(monkeypatch, FakeChatModel)
+
+
+@pytest.fixture
+def stub_failing_llamafactory(monkeypatch) -> None:
+    _install_llamafactory(monkeypatch, FailingChatModel)
+
+
+def _wait_for_status(session: InferenceSession, status: str, timeout: float = 3.0) -> bool:
+    deadline = time.time() + timeout
+    while time.time() < deadline:
+        if session.info()["status"] == status:
+            return True
+        time.sleep(0.02)
+    return False
+
+
+def test_load_returns_immediately_then_ready(stub_llamafactory) -> None:
+    session = InferenceSession()
+    started = time.time()
+    result = session.load("/models/qwen")
+    assert result["status"] == "loading"
+    assert result["loaded"] is False
+    assert result["request_id"]
+    # 在慢加载完成前就返回，且 info() 加载期间可响应
+    assert time.time() - started < LOAD_DELAY
+    assert session.info()["status"] == "loading"
+    assert _wait_for_status(session, "ready")
+    info = session.info()
+    assert info["loaded"] is True
+    assert info["status"] == "ready"
+    assert info["model_name"] == "/models/qwen"
+
+
+def test_second_load_while_loading_deduped(stub_llamafactory) -> None:
+    session = InferenceSession()
+    r1 = session.load("/models/a")
+    r2 = session.load("/models/b")
+    assert r2["status"] == "loading"
+    assert r2["request_id"] == r1["request_id"]
+    assert _wait_for_status(session, "ready")
+    assert session.info()["status"] == "ready"
+
+
+def test_load_error_surfaces_in_status(stub_failing_llamafactory) -> None:
+    session = InferenceSession()
+    session.load("/models/bad")
+    assert _wait_for_status(session, "error")
+    assert "boom" in session.info()["error"]
+
+
+def test_unload_while_loading_cancels(stub_llamafactory) -> None:
+    session = InferenceSession()
+    session.load("/models/qwen")
+    result = session.unload()
+    assert result["status"] == "cancelling"
+    assert _wait_for_status(session, "idle")
+
+
+def test_chat_while_loading_returns_loading_error(stub_llamafactory) -> None:
+    session = InferenceSession()
+    session.load("/models/qwen")
+    out = session.chat([{"role": "user", "content": "hi"}])
+    assert "still loading" in (out.get("error") or "")
+    assert _wait_for_status(session, "ready")
+    out = session.chat([{"role": "user", "content": "hi"}])
+    assert out.get("response") == "hello from model"
+
+
+def test_chat_stream_while_loading_yields_error(stub_llamafactory) -> None:
+    session = InferenceSession()
+    session.load("/models/qwen")
+    chunks = list(session.chat_stream([{"role": "user", "content": "hi"}]))
+    assert any("still loading" in c for c in chunks)
+
+
+def test_wait_until_loaded_blocks_until_ready(stub_llamafactory) -> None:
+    session = InferenceSession()
+    result = session.load("/models/qwen")
+    assert result["status"] == "loading"
+    # 同步等待后台加载线程完成
+    outcome = session.wait_until_loaded(timeout=3.0)
+    assert outcome["loaded"] is True
+    assert outcome["status"] == "ready"
+
+
+def test_wait_until_loaded_reports_load_error(stub_failing_llamafactory) -> None:
+    session = InferenceSession()
+    session.load("/models/bad")
+    outcome = session.wait_until_loaded(timeout=3.0)
+    assert outcome["loaded"] is False
+    assert outcome["status"] == "error"
+    assert "boom" in outcome["error"]

docker/nginx.conf.template

@@ -14,8 +14,8 @@ server {
     proxy_set_header X-Real-IP $remote_addr;
     proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
     proxy_set_header X-Forwarded-Proto $scheme;
-    proxy_read_timeout 300s;
-    proxy_send_timeout 300s;
+    proxy_read_timeout 900s;
+    proxy_send_timeout 900s;
   }
 
   location = /modelTF {
@@ -25,8 +25,8 @@ server {
     proxy_set_header X-Real-IP $remote_addr;
     proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
     proxy_set_header X-Forwarded-Proto $scheme;
-    proxy_read_timeout 300s;
-    proxy_send_timeout 300s;
+    proxy_read_timeout 900s;
+    proxy_send_timeout 900s;
   }
 
   location ~* \.(?:js|css|png|jpg|jpeg|gif|ico|svg|woff|woff2|ttf)$ {

frontend/src/api/modules/audit-visit.ts

@@ -1,5 +0,0 @@
-import { post } from '../request'
-
-/** 记录用户访问某个业务模块（用于看板用户操作分布统计） */
-export const recordModuleVisit = (module: string, detail?: string) =>
-  post('/system/audit/visit', { action: module, detail: detail || '' })

frontend/src/api/modules/compare.ts

@@ -1,6 +1,8 @@
 import { get, post, del } from '../request'
 import type { CompareTask, CompareModelRef } from '@/types'
 
+const INFERENCE_START_TIMEOUT_MS = 15 * 60 * 1000
+
 /** 推理/对比任务列表 */
 export const getCompareList = () => get<CompareTask[]>('/model-compare')
 
@@ -12,7 +14,7 @@ export const createCompare = (data: Partial<CompareTask>) =>
   post<{ id: string | number }>('/model-compare', data)
 
 /** 删除任务 */
-export const deleteCompare = (id: string | number) => del(`/model-compare/${id}`)
+export const deleteCompare = (id: string | number) => del(`/model-compare/${id}`, undefined, { timeout: 60_000 })
 
 /** 更新任务加载状态 */
 export const updateLoadStatus = (id: string | number, load_status: any) =>
@@ -34,7 +36,8 @@ export const stopModelByPid = (pid: number) =>
   post('/model-compare/stop-by-pid', { pid })
 
 /** 加载任务 */
-export const loadCompare = (id: string | number) => post(`/model-compare/${id}/load`)
+export const loadCompare = (id: string | number) =>
+  post(`/model-compare/${id}/load`, undefined, { timeout: INFERENCE_START_TIMEOUT_MS })
 
 /** 卸载任务 */
 export const unloadCompare = (id: string | number) => post(`/model-compare/${id}/unload`)
@@ -81,6 +84,10 @@ export const streamChatReal = (data: any): Promise<Response> => {
       temperature: data.temperature ?? 0.7,
       top_p: data.top_p ?? 0.95,
       max_tokens: data.max_tokens ?? 2048,
+      // 透传 task_id/node_id，让后端按 load_status 路由到真正加载了模型的算力节点，
+      // 避免在多节点时回退到“第一个在线节点”导致连接失败
+      task_id: data.task_id,
+      node_id: data.node_id,
     }),
   })
 }
@@ -94,8 +101,8 @@ export const batchChat = (data: any) => post('/model-chat/batch', data)
 /** 本地 transformers 模型对话 */
 export const localChat = (data: any) => post('/model-chat/local/chat', data)
 
-/** 预加载本地模型（模型加载耗时长，超时 5 分钟） */
-export const preloadLocalModel = (data: any) => post('/model-chat/local/preload', data, { timeout: 300000 })
+/** 预加载本地模型（模型加载耗时长，超时 15 分钟） */
+export const preloadLocalModel = (data: any) => post('/model-chat/local/preload', data, { timeout: INFERENCE_START_TIMEOUT_MS })
 
-/** 预加载已训练模型（超时 5 分钟） */
-export const preloadTrainedModel = (data: any) => post('/model-chat/trained/preload', data, { timeout: 300000 })
+/** 预加载已训练模型（超时 15 分钟） */
+export const preloadTrainedModel = (data: any) => post('/model-chat/trained/preload', data, { timeout: INFERENCE_START_TIMEOUT_MS })

frontend/src/api/modules/fineTune.ts

@@ -1,6 +1,16 @@
 import { get, post, put, del } from '../request'
 import type { FineTuneStartPayload, FineTuneTask, TrainingProgress, LogContent } from '@/types'
 
+export interface FineTuneMetricPoint {
+  step: number
+  epoch?: number | null
+  loss?: number | null
+  grad_norm?: number | null
+  learning_rate?: number | null
+  raw?: string
+  create_time?: string
+}
+
 export interface TrainingDiagnostic {
   level: string
   title: string
@@ -80,6 +90,10 @@ export const getFineTuneLogs = (
   params: { tail_lines?: number; offset?: number; limit?: number } = {},
 ) => get<LogContent & { job_id?: string; source?: string }>(`/fine-tune/${id}/logs`, params)
 
+/** 获取训练指标曲线数据 */
+export const getFineTuneMetrics = (id: string | number) =>
+  get<FineTuneMetricPoint[]>(`/fine-tune/${id}/metrics`)
+
 /** 启动 TensorBoard */
 export const startTensorboard = () => post('/fine-tune/tensorboard/start')
 

frontend/src/api/modules/model.ts

@@ -88,10 +88,14 @@ export const updateModelPurpose = (id: string | number, purpose: string) =>
 
 /** 合并 LoRA 权重 */
 export const mergeModel = (data: {
+  trained_model_id?: string | number
   model_name: string
   train_method: string
   base_model_path: string
-}) => post('/model-manage/merge', data)
+  adapter_path?: string
+  compute_node_id?: string
+  output_model_name?: string
+}) => post('/model-manage/merge', data, { timeout: 15 * 60 * 1000 })
 
 /** 导出已训练模型权重 */
 export const exportModelUrl = (modelName: string) =>

frontend/src/api/modules/system.ts

@@ -8,6 +8,8 @@ import type {
   UpdateUserAccessPayload,
 } from '@/types'
 
+export type { SystemUser } from '@/types'
+
 /** 系统信息（CPU/内存/磁盘/GPU/网络/系统） */
 export const getSystemInfo = () => get<SystemInfo>('/system-info')
 

frontend/src/api/request.ts

@@ -1,37 +1,6 @@
 import axios, { type AxiosInstance, type AxiosRequestConfig } from 'axios'
 import { ElMessage } from 'element-plus'
 
-/**
- * 用户操作分布：哪些模块路径算"业务操作"（用于看板统计）
- * 请求命中这些路径时，会自动调用 record_visit 记录一次（同一模块 60 秒内去重）
- */
-const VISIT_TRACKED_PREFIXES: Array<[string, string]> = [
-  ['/fine-tune', 'fine-tune'],
-  ['/model-eval', 'model-eval'],
-  ['/model-inference', 'model-inference'],
-  ['/data-process', 'data-process'],
-  ['/data-convert', 'data-convert'],
-  ['/model-manage', 'model-manage'],
-  ['/dataset-manage', 'dataset'],
-]
-
-function trackVisit(url: string | undefined) {
-  if (!url) return
-  for (const [prefix, module] of VISIT_TRACKED_PREFIXES) {
-    if (url.includes(prefix)) {
-      const key = `visit:${module}`
-      const last = Number(sessionStorage.getItem(key) || 0)
-      if (Date.now() - last < 60000) return // 60 秒内去重
-      sessionStorage.setItem(key, String(Date.now()))
-      // fire-and-forget 调用后端记录接口
-      import('./modules/audit-visit').then(({ recordModuleVisit }) => {
-        recordModuleVisit(module, url).catch(() => { /* ignore */ })
-      }).catch(() => { /* ignore */ })
-      return
-    }
-  }
-}
-
 /**
  * 后端统一响应格式
  * code === 0 表示成功，data 为业务数据
@@ -91,8 +60,6 @@ service.interceptors.response.use(
       return response
     }
     if (res.code === 0) {
-      // 记录业务模块访问（用于看板用户操作分布统计）
-      trackVisit(response.config.url)
       return res.data
     }
     // 业务错误

frontend/src/composables/useStreamChat.ts

@@ -44,6 +44,24 @@ export function useStreamChat() {
   })
   const loading = ref(false)
 
+  /** 从 SSE 帧中提取错误信息（后端/计算节点错误以 data: {"error": "..."} 形式下发） */
+  function extractSseError(buffer: string): string | null {
+    const trimmed = buffer.trim()
+    if (!trimmed.startsWith('data: ')) return null
+    const lines = trimmed.split(/\r?\n/)
+    for (let i = lines.length - 1; i >= 0; i--) {
+      const line = lines[i].trim()
+      if (!line.startsWith('data: ')) continue
+      try {
+        const obj = JSON.parse(line.slice(6))
+        if (obj && typeof obj.error === 'string' && obj.error) return obj.error
+      } catch {
+        /* 非 JSON 的 data 行忽略 */
+      }
+    }
+    return trimmed
+  }
+
   /** 从内容中解析 think 标签 */
   function parseContent(content: string) {
     const thinkRegex = /<think>([\s\S]*?)(<\/think>)?/g
@@ -121,6 +139,16 @@ export function useStreamChat() {
       }
 
       // 最终更新
+      // 若整段响应是 SSE 错误帧，提取 error 字段以干净文案展示
+      const sseError = extractSseError(buffer)
+      if (sseError) {
+        message.value.isThinking = false
+        message.value.isStreaming = false
+        message.value.done = true
+        message.value.error = sseError
+        message.value.displayContent = sseError
+        return
+      }
       const parsed = parseContent(buffer)
       message.value.thinkContent = parsed.think
       message.value.displayContent = parsed.display

frontend/src/types/index.ts

@@ -34,6 +34,10 @@ export interface TrainedModel {
   name: string
   train_methods?: TrainMethod[]
   base_model_path?: string
+  artifact_dir?: string
+  adapter_path?: string
+  compute_node_id?: string
+  compute_node_name?: string
   create_time?: string
   merged?: boolean
   merging?: boolean
@@ -212,6 +216,9 @@ export interface LoadedModel {
   status?: string
   pid?: number
   port?: number
+  node_id?: string
+  node_name?: string
+  error?: string
 }
 
 export interface CompareTask {
@@ -230,6 +237,8 @@ export interface CompareModelRef {
   model_name: string
   model_path: string
   gpu_id: number
+  node_id?: string
+  node_name?: string
   source?: string
   port?: number
 }
@@ -245,7 +254,9 @@ export interface EvalTask {
   model_name?: string
   model_id?: number | string
   dataset?: string
+  dataset_id?: number | string
   metric?: string
+  metric_label?: string
   score?: number
   status?: string
   create_time?: string
@@ -272,6 +283,7 @@ export interface StartEvalPayload {
   eval_type: EvalType
   model_id: string | number
   gpu_id: string | number
+  compute_node_id?: string
   dataset_id: string | number
   dimension_id: string | number
   data_source: 'dataset' | 'inference'

frontend/src/views/dashboard/DashboardView.vue

@@ -66,8 +66,12 @@ const operationDistribution = ref<{ name: string; value: number }[]>([])
 const serviceIcon: Record<string, string> = {
   '模型推理': 'fa-cube',
   '模型微调': 'fa-sliders',
+  '模型训练': 'fa-sliders',
   '模型评测': 'fa-bar-chart',
+  '模型管理': 'fa-cubes',
+  '数据集管理': 'fa-file-text',
   '数据处理': 'fa-filter',
+  '数据类型转换': 'fa-exchange',
 }
 const roleLabel: Record<string, string> = {
   admin: '超级管理员',
@@ -115,7 +119,7 @@ const chartOption = computed<EChartsOption>(() => ({
     borderWidth: 0,
     padding: [10, 12],
     textStyle: { color: '#ffffff', fontSize: 12 },
-    valueFormatter: (value) => `${value}`,
+    valueFormatter: (value) => String(value ?? ''),
   },
   xAxis: {
     type: 'category',
@@ -252,7 +256,7 @@ const loginDurationChartOption = computed<EChartsOption>(() => {
     tooltip: {
       trigger: 'axis',
       axisPointer: { type: 'shadow' },
-      valueFormatter: (value: number) => `${value} 小时`,
+      valueFormatter: (value: unknown) => String(Number(Array.isArray(value) ? value[0] : value) || 0) + ' 小时',
     },
     xAxis: {
       type: 'value',
@@ -738,10 +742,11 @@ function viewTask(task: DashboardTask) {
 
 .service-table {
   display: grid;
-  grid-template-rows: 36px repeat(4, minmax(48px, 1fr));
+  grid-auto-rows: minmax(44px, auto);
   flex: 1 1 auto;
   margin-top: 12px;
   min-height: 0;
+  overflow-y: auto;
 }
 
 .service-row {
@@ -935,7 +940,7 @@ function viewTask(task: DashboardTask) {
   }
 
   .service-table {
-    grid-template-rows: 32px repeat(4, minmax(40px, 1fr));
+    grid-auto-rows: minmax(38px, auto);
     margin-top: 8px;
   }
 

frontend/src/views/eval/EvalCreateView.vue

@@ -1,4 +1,4 @@
-<script setup lang="ts">
+<script setup lang="ts">
 import { onMounted, ref, watch } from 'vue'
 import { useRouter } from 'vue-router'
 import { ElMessage } from 'element-plus'
@@ -10,8 +10,7 @@ import StartEvalStep from './create/StartEvalStep.vue'
 import { createDimension, startEval } from '@/api/modules/eval'
 import { getTrainedModels, getModelList } from '@/api/modules/model'
 import { getDatasetList } from '@/api/modules/dataset'
-import { getSystemInfo } from '@/api/modules/system'
-import { getComputeNodes, type ComputeNode } from '@/api/modules/compute'
+import { getComputeGpus } from '@/api/modules/compute'
 import type { DatasetItem, Dimension, GpuInfo, ModelItem, TrainedModel } from '@/types'
 
 type StepExposed = { validate: () => Promise<boolean> }
@@ -83,9 +82,8 @@ async function loadData() {
   const results = await Promise.allSettled([
     getTrainedModels(),
     getDatasetList(),
-    getSystemInfo(),
     getModelList(),
-    getComputeNodes(),
+    getComputeGpus(),
   ])
 
   if (results[0].status === 'fulfilled') trainedModels.value = results[0].value?.models || []
@@ -93,18 +91,12 @@ async function loadData() {
     evalDatasets.value = (results[1].value || []).filter((dataset) => dataset.type === 'eval')
   }
   if (results[2].status === 'fulfilled') {
-    const allGpus: GpuInfo[] = results[2].value?.gpu || []
-    const nodes: ComputeNode[] = (results[4].status === 'fulfilled' ? results[4].value : []) || []
-    const onlineIds = new Set(nodes.filter((n) => n.enabled && n.scheduler_status === 'online').map((n) => n.id))
-    // Only show idle GPUs from online compute nodes
-    gpus.value = allGpus.filter(
-      (g) => g.status === 'idle' && (!g.node_id || onlineIds.has(g.node_id)),
+    evalModels.value = (results[2].value || []).filter(
+      (model) => model.purpose === 'evaluation' || (model.model_source === 'api' && !!model.api_url),
     )
   }
   if (results[3].status === 'fulfilled') {
-    evalModels.value = (results[3].value || []).filter(
-      (model) => model.purpose === 'evaluation' || (model.model_source === 'api' && !!model.api_url),
-    )
+    gpus.value = ((results[3].value || []) as unknown as GpuInfo[]).filter((g) => g.status === 'idle')
   }
 
   const failedCount = results.filter((result) => result.status === 'rejected').length
@@ -151,11 +143,15 @@ async function handleSubmit() {
   submitting.value = true
   try {
     const dimensionId = await resolveDimensionId()
-    await startEval({
+    // GPU 选择为「节点:GPU序号」复合值，解析出节点与 GPU 序号，
+    // 多算力节点时必须把节点信息传给后端，否则会派发到错误的算力节点
+    const [gpuNodeId, gpuIndex] = String(taskForm.value.gpu_id).split(':')
+    const evalResult: any = await startEval({
       eval_task_name: taskForm.value.eval_task_name,
       eval_type: 'custom',
       model_id: taskForm.value.model_id,
-      gpu_id: taskForm.value.gpu_id,
+      gpu_id: Number(gpuIndex) || 0,
+      compute_node_id: gpuNodeId || '',
       dataset_id: taskForm.value.data_source === 'dataset' ? taskForm.value.dataset_id : '',
       dimension_id: dimensionId,
       data_source: taskForm.value.data_source,
@@ -175,6 +171,10 @@ async function handleSubmit() {
         output_precision: basicMetricForm.value.output_precision,
       },
     })
+    if (evalResult?.status === 'failed' || evalResult?.error) {
+      ElMessage.error(`评测启动失败：${evalResult?.error || '请检查算力节点与模型路径'}`)
+      return
+    }
     ElMessage.success('评测任务已创建并启动')
     router.push('/model-eval')
   } catch (error) {

frontend/src/views/eval/EvalDetailView.vue

@@ -1,4 +1,4 @@
-<script setup lang="ts">
+<script setup lang="ts">
 import { computed, onMounted, onUnmounted, ref } from 'vue'
 import { useRoute } from 'vue-router'
 import PageCard from '@/components/PageCard.vue'
@@ -50,6 +50,8 @@ const passRate = computed(() => {
 })
 
 const overallScore = computed(() => formatScore(detail.value?.overall_score, detail.value?.overall_score_max))
+const displayModelName = computed(() => detail.value?.model_name || String(detail.value?.model_id || '-'))
+const displayMetric = computed(() => detail.value?.metric_label || detail.value?.metric || '-')
 
 function formatDateTime(value?: string) {
   if (!value) return '-'
@@ -121,9 +123,9 @@ onUnmounted(stopPolling)
           </div>
           <dl class="task-meta">
             <div><dt>任务 ID</dt><dd>{{ detail?.id || taskId }}</dd></div>
-            <div><dt>评测模型</dt><dd>{{ detail?.model_name || '-' }}</dd></div>
+            <div><dt>评测模型</dt><dd>{{ displayModelName }}</dd></div>
             <div><dt>测试集</dt><dd>{{ detail?.dataset || '-' }}</dd></div>
-            <div><dt>评测指标</dt><dd>{{ detail?.metric || '-' }}</dd></div>
+            <div><dt>评测指标</dt><dd>{{ displayMetric }}</dd></div>
           </dl>
         </div>
       </div>
@@ -141,7 +143,7 @@ onUnmounted(stopPolling)
         <div class="overview-item score-hero">
           <span>综合得分</span>
           <strong>{{ overallScore }}</strong>
-          <small>大模型综合评分</small>
+          <small>模型综合评分</small>
         </div>
         <div class="overview-item">
           <span>样本通过率</span>
@@ -164,7 +166,7 @@ onUnmounted(stopPolling)
         <div class="review-copy">
           <div class="section-heading">
             <div>
-              <h2 id="overall-review-title">大模型综合评价</h2>
+              <h2 id="overall-review-title">综合评价</h2>
               <p>基于全部已评测样本生成的总体结论</p>
             </div>
             <el-tag v-if="detail.evaluator_model" type="primary" size="small">
@@ -172,7 +174,7 @@ onUnmounted(stopPolling)
             </el-tag>
           </div>
           <p class="review-text">
-            {{ detail.overall_evaluation || (detail.status === 'running' ? '评测仍在进行，综合评价将在样本评分完成后生成。' : '暂无综合评价。') }}
+            {{ detail.overall_evaluation || (detail.status === 'running' ? '评测正在进行，综合评价将在样本完成后生成。' : '暂无综合评价。') }}
           </p>
 
           <div class="suggestion-block">
@@ -190,8 +192,8 @@ onUnmounted(stopPolling)
       <section v-if="detail.dimension_summary?.length" class="dimension-summary" aria-labelledby="dimension-title">
         <div class="section-heading compact-heading">
           <div>
-            <h2 id="dimension-title">维度表现</h2>
-            <p>查看各评测维度的得分与样本通过率</p>
+            <h2 id="dimension-title">指标表现</h2>
+            <p>查看各评测指标的得分与通过率</p>
           </div>
         </div>
         <div class="dimension-grid">
@@ -212,22 +214,10 @@ onUnmounted(stopPolling)
             <p>共 {{ filteredSamples.length }} 条结果，展开行可查看评分依据与子维度分数</p>
           </div>
           <div class="sample-filters" aria-label="样本筛选">
-            <el-input
-              v-model="keyword"
-              clearable
-              placeholder="搜索问题、回答或评价"
-              aria-label="搜索样本"
-              @input="resetPage"
-            >
+            <el-input v-model="keyword" clearable placeholder="搜索问题、回答或评价" aria-label="搜索样本" @input="resetPage">
               <template #prefix><i class="fa fa-search" aria-hidden="true" /></template>
             </el-input>
-            <el-select
-              v-model="judgementFilter"
-              clearable
-              placeholder="全部判定"
-              aria-label="按判定筛选"
-              @change="resetPage"
-            >
+            <el-select v-model="judgementFilter" clearable placeholder="全部判定" aria-label="按判定筛选" @change="resetPage">
               <el-option label="正确" value="正确" />
               <el-option label="部分正确" value="部分正确" />
               <el-option label="错误" value="错误" />
@@ -235,18 +225,12 @@ onUnmounted(stopPolling)
           </div>
         </div>
 
-        <el-table
-          v-if="filteredSamples.length"
-          class="sample-results-table"
-          :data="paginatedSamples"
-          row-key="id"
-          table-layout="fixed"
-        >
+        <el-table v-if="filteredSamples.length" class="sample-results-table" :data="paginatedSamples" row-key="id" table-layout="fixed">
           <el-table-column type="expand" width="48">
             <template #default="{ row }">
               <div class="sample-detail-grid">
                 <div class="evaluation-reason">
-                  <span>大模型评分依据</span>
+                  <span>评分依据</span>
                   <p>{{ row.evaluation_reason || '暂无评分依据。' }}</p>
                 </div>
                 <div v-if="row.error_type" class="error-type">
@@ -275,9 +259,7 @@ onUnmounted(stopPolling)
             <template #default="{ row }"><p class="cell-copy">{{ row.model_output || '等待生成' }}</p></template>
           </el-table-column>
           <el-table-column label="得分" width="90" align="center">
-            <template #default="{ row }">
-              <span class="sample-score">{{ formatScore(row.score, row.max_score) }}</span>
-            </template>
+            <template #default="{ row }"><span class="sample-score">{{ formatScore(row.score, row.max_score) }}</span></template>
           </el-table-column>
           <el-table-column label="判定" width="96" align="center">
             <template #default="{ row }">
@@ -294,21 +276,11 @@ onUnmounted(stopPolling)
           <p>{{ detail.status === 'running' ? '任务正在运行，结果生成后会显示在这里。' : '请调整筛选条件或稍后重试。' }}</p>
         </div>
 
-        <el-pagination
-          v-if="filteredSamples.length > pageSize"
-          v-model:current-page="currentPage"
-          v-model:page-size="pageSize"
-          background
-          layout="total, sizes, prev, pager, next"
-          :page-sizes="[10, 20, 50]"
-          :total="filteredSamples.length"
-          aria-label="样本结果分页"
-        />
+        <el-pagination v-if="filteredSamples.length > pageSize" v-model:current-page="currentPage" v-model:page-size="pageSize" background layout="total, sizes, prev, pager, next" :page-sizes="[10, 20, 50]" :total="filteredSamples.length" aria-label="样本结果分页" />
       </section>
     </template>
   </PageCard>
 </template>
-
 <style scoped lang="scss">
 .eval-detail-page {
   min-width: 0;
@@ -767,3 +739,6 @@ onUnmounted(stopPolling)
   }
 }
 </style>
+
+
+

frontend/src/views/eval/EvalView.vue

@@ -1,4 +1,4 @@
-<script setup lang="ts">
+<script setup lang="ts">
 import { ref, onMounted, onUnmounted } from 'vue'
 import { useRouter } from 'vue-router'
 import { ElMessage, ElMessageBox } from 'element-plus'
@@ -57,6 +57,14 @@ function handleViewDetail(row: any) {
   router.push({ name: 'model-eval-detail', params: { id: row.id } })
 }
 
+function displayModelName(row: Partial<EvalTask>) {
+  return row.model_name || String(row.model_id || '-')
+}
+
+function displayMetric(row: Partial<EvalTask>) {
+  return row.metric_label || row.metric || '-'
+}
+
 const { start: startPolling, stop: stopPolling } = usePolling(
   async () => {
     await loadEvalList({ silent: true })
@@ -102,9 +110,21 @@ onUnmounted(() => {
       </template>
             <template #columns>
               <el-table-column label="任务名称" prop="eval_task_name" align="center" />
-              <el-table-column label="评测模型" prop="model_name" align="center" />
+              <el-table-column label="评测模型" align="center" min-width="160">
+                <template #default="{ row }">
+                  <el-tooltip :content="displayModelName(row)" placement="top" :disabled="displayModelName(row).length < 18">
+                    <span class="cell-ellipsis">{{ displayModelName(row) }}</span>
+                  </el-tooltip>
+                </template>
+              </el-table-column>
               <el-table-column label="数据集" prop="dataset" align="center" />
-              <el-table-column label="指标" prop="metric" align="center" />
+              <el-table-column label="指标" align="center" min-width="220">
+                <template #default="{ row }">
+                  <el-tooltip :content="displayMetric(row)" placement="top" :disabled="displayMetric(row).length < 24">
+                    <span class="cell-ellipsis">{{ displayMetric(row) }}</span>
+                  </el-tooltip>
+                </template>
+              </el-table-column>
               <el-table-column label="评分" prop="score" width="100" align="center" />
               <el-table-column label="状态" width="100" align="center">
                 <template #default="{ row }">
@@ -159,7 +179,7 @@ onUnmounted(() => {
   min-height: 0;
 }
 
-/* 胶囊切换栏样式 */
+/* 胶囊切换栏 */
 .capsule-tabs {
   display: flex;
   background: #f1f5f9;
@@ -193,4 +213,13 @@ onUnmounted(() => {
   }
 }
 
+.cell-ellipsis {
+  display: inline-block;
+  max-width: 100%;
+  overflow: hidden;
+  text-overflow: ellipsis;
+  vertical-align: middle;
+  white-space: nowrap;
+}
+
 </style>

frontend/src/views/eval/create/EvalTaskSetupStep.vue

@@ -104,9 +104,9 @@ defineExpose({ validate })
       <el-select v-model="form.gpu_id" placeholder="请选择 GPU" style="width: 100%" :loading="loading">
         <el-option
           v-for="gpu in gpus"
-          :key="gpu.id"
-          :label="`${gpu.name} (GPU ${gpu.id})`"
-          :value="gpu.id ?? 0"
+          :key="`${gpu.node_id || ''}:${gpu.id ?? 0}`"
+          :label="`${gpu.node_name || gpu.node_code || '算力节点'} / ${gpu.name} (GPU ${gpu.id ?? 0})`"
+          :value="`${gpu.node_id || ''}:${gpu.id ?? 0}`"
         />
       </el-select>
     </el-form-item>

frontend/src/views/eval/create/StartEvalStep.vue

@@ -1,4 +1,5 @@
 <script setup lang="ts">
+import { computed } from 'vue'
 import type { DatasetItem, GpuInfo, ModelItem, TrainedModel } from '@/types'
 import type { BasicMetricSetupDraft } from './BasicMetricSetupStep.vue'
 import type { EvalRuleSetupDraft } from './EvalRuleSetupStep.vue'
@@ -17,6 +18,15 @@ const props = defineProps<{
 function nameOf<T extends { id: string | number; name?: string }>(items: T[], id: string | number) {
   return items.find((item) => item.id === id)?.name || String(id || '-')
 }
+
+/** GPU 选择为「节点:GPU序号」复合值，解析并展示为可读标签 */
+const gpuLabel = computed(() => {
+  const key = String(props.task.gpu_id || '')
+  const gpu = props.gpus.find((g) => `${g.node_id || ''}:${g.id ?? 0}` === key)
+  if (gpu) return `${gpu.node_name || gpu.node_code || '算力节点'} / GPU ${gpu.id ?? 0}`
+  const [nodeId, idx] = key.split(':')
+  return nodeId ? `节点 ${nodeId} / GPU ${idx || 0}` : `GPU ${key || 0}`
+})
 </script>
 
 <template>
@@ -24,7 +34,7 @@ function nameOf<T extends { id: string | number; name?: string }>(items: T[], id
     <el-descriptions :column="2" border>
       <el-descriptions-item label="Task">{{ props.task.eval_task_name || '-' }}</el-descriptions-item>
       <el-descriptions-item label="Model">{{ nameOf(props.trainedModels, props.task.model_id) }}</el-descriptions-item>
-      <el-descriptions-item label="GPU">GPU {{ props.task.gpu_id || 0 }}</el-descriptions-item>
+      <el-descriptions-item label="GPU">{{ gpuLabel }}</el-descriptions-item>
       <el-descriptions-item label="Dataset">
         {{ props.task.data_source === 'dataset' ? nameOf(props.evalDatasets, props.task.dataset_id) : 'Inference results' }}
       </el-descriptions-item>

frontend/src/views/inference/InferenceChatView.vue

@@ -1,10 +1,10 @@
 <script setup lang="ts">
-import { ref, reactive, nextTick, onMounted, watch } from 'vue'
+import { ref, reactive, nextTick, onMounted, onUnmounted, watch } from 'vue'
 import { useRoute, useRouter } from 'vue-router'
 import { ElMessage } from 'element-plus'
 import MarkdownView from '@/components/MarkdownView.vue'
 import { useStreamChat } from '@/composables/useStreamChat'
-import { getCompare } from '@/api/modules/compare'
+import { getCompare, getLoadStatus } from '@/api/modules/compare'
 import type { CompareTask, LoadedModel } from '@/types'
 
 const route = useRoute()
@@ -37,6 +37,10 @@ const contentRef = ref<HTMLElement>()
 let activeAssistant: ChatMessage | null = null
 /** 设置面板抽屉 */
 const showSettings = ref(false)
+/** 模型仍在加载中（直接 URL 进入 chat 时兜底轮询就绪状态） */
+const taskLoading = ref(false)
+const taskError = ref('')
+let statusTimer: ReturnType<typeof setInterval> | null = null
 
 /** 获取任务信息，定位已启动的模型（mock 模式跳过） */
 async function loadTask() {
@@ -45,6 +49,14 @@ async function loadTask() {
     task.value = await getCompare(taskId)
     const models = parseLoadedModels(task.value)
     if (models[0]?.model_name) modelName.value = models[0].model_name
+    // 恢复本地保存的历史对话
+    restoreHistory()
+    // 模型仍在上次加载中：启动轮询等待就绪
+    if (models.some((m) => m.status === 'starting')) {
+      taskLoading.value = true
+      await pollTaskStatus()
+      statusTimer = setInterval(pollTaskStatus, 3000)
+    }
   } catch {
     // ignore
   }
@@ -60,6 +72,80 @@ function parseLoadedModels(t: CompareTask | null): LoadedModel[] {
   }
 }
 
+/** 对话历史本地持久化（按任务 id 存储，退出重进可恢复） */
+const STORAGE_PREFIX = 'ygft_chat_history_'
+
+function historyKey(id: string | number): string {
+  return `${STORAGE_PREFIX}${id}`
+}
+
+function saveHistory() {
+  if (isMock) return
+  try {
+    const snapshot = messages.value.map((m) => ({
+      role: m.role,
+      content: m.content,
+      think: m.think,
+      done: true,
+    }))
+    localStorage.setItem(historyKey(taskId), JSON.stringify(snapshot))
+  } catch {
+    // 存储失败忽略
+  }
+}
+
+function restoreHistory() {
+  if (isMock) return
+  try {
+    const raw = localStorage.getItem(historyKey(taskId))
+    if (!raw) return
+    const parsed = JSON.parse(raw)
+    if (Array.isArray(parsed)) {
+      messages.value = parsed.map((m) => ({
+        role: m.role === 'user' ? 'user' : 'assistant',
+        content: m.content || '',
+        think: m.think || '',
+        isThinking: false,
+        isStreaming: false,
+        done: true,
+      }))
+    }
+  } catch {
+    // 恢复失败忽略
+  }
+}
+
+/** 停止就绪状态轮询 */
+function stopStatusPolling() {
+  if (statusTimer) {
+    clearInterval(statusTimer)
+    statusTimer = null
+  }
+}
+
+/** 轮询任务加载状态：starting → ready/error */
+async function pollTaskStatus() {
+  try {
+    const st = await getLoadStatus(taskId)
+    const items = st.loaded_models || []
+    const anyReady = items.some((m) => m.status === 'ready' || m.status === 'running')
+    const anyError = items.some((m) => m.status === 'error')
+    if (anyReady) {
+      taskLoading.value = false
+      taskError.value = ''
+      stopStatusPolling()
+    } else if (anyError) {
+      taskLoading.value = false
+      taskError.value = items.find((m) => m.status === 'error')?.error || '模型加载失败'
+      stopStatusPolling()
+    } else {
+      taskLoading.value = true
+    }
+  } catch {
+    // 轮询失败忽略，下次再试
+  }
+}
+
 async function handleSend() {
   const question = inputQuestion.value.trim()
   if (!question || loading.value) return
@@ -76,6 +162,7 @@ async function handleSend() {
     done: false,
   })
   messages.value.push(assistantMsg)
+  saveHistory()
 
   inputQuestion.value = ''
   await nextTick()
@@ -85,6 +172,7 @@ async function handleSend() {
   // mock 模式：直接用假数据逐字填充
   if (isMock) {
     await mockReply(assistantMsg, question)
+    saveHistory()
     return
   }
 
@@ -94,6 +182,7 @@ async function handleSend() {
   await send(
     {
       model_path: route.query.model_path as string || '',
+      task_id: taskId,
       system_prompt: systemPrompt.value,
       user_question: question,
       temperature: temperature.value,
@@ -111,6 +200,7 @@ async function handleSend() {
   assistantMsg.done = true
   activeAssistant = null
   reset()
+  saveHistory()
   await nextTick()
   scrollToBottom()
 }
@@ -169,6 +259,11 @@ function handleNewChat() {
   activeAssistant = null
   messages.value = []
   reset()
+  try {
+    localStorage.removeItem(historyKey(taskId))
+  } catch {
+    // 忽略
+  }
 }
 
 /** 输入框自适应高度 */
@@ -185,6 +280,7 @@ function resetInputHeight() {
 }
 
 onMounted(loadTask)
+onUnmounted(stopStatusPolling)
 </script>
 
 <template>
@@ -252,6 +348,12 @@ onMounted(loadTask)
 
     <!-- 输入栏 -->
     <footer class="chat-input-container">
+      <div v-if="taskLoading" class="loading-hint">
+        <i class="fa fa-spinner fa-spin" style="margin-right: 6px" />模型加载中，就绪后即可对话...
+      </div>
+      <div v-else-if="taskError" class="loading-hint error">
+        <i class="fa fa-exclamation-circle" style="margin-right: 6px" />{{ taskError }}
+      </div>
       <div class="chat-input-inner">
         <button class="clear-btn" title="清空对话" @click="handleNewChat">
           <i class="fa fa-eraser" />
@@ -261,22 +363,21 @@ onMounted(loadTask)
             v-model="inputQuestion"
             class="input-box"
             rows="1"
-            :disabled="loading"
+            :disabled="loading || taskLoading"
             placeholder="给模型发送消息..."
             @keydown.enter.exact.prevent="handleSend"
             @input="autoResize"
           />
           <button
             class="send-btn"
-            :class="{ active: inputQuestion.trim() && !loading }"
-            :disabled="!inputQuestion.trim() || loading"
+            :class="{ active: inputQuestion.trim() && !loading && !taskLoading }"
+            :disabled="!inputQuestion.trim() || loading || taskLoading"
             @click="handleSend"
           >
             <i class="fa fa-arrow-up" />
           </button>
         </div>
       </div>
-      <div class="footer-hint">内容由 AI 生成，请仔细甄别。</div>
     </footer>
 
     <!-- 设置抽屉（系统提示词等） -->
@@ -703,9 +804,18 @@ onMounted(loadTask)
   }
 }
 
-.footer-hint {
-  margin-top: 12px;
-  font-size: 12px;
-  color: #9ca3af;
+.loading-hint {
+  margin-bottom: 10px;
+  padding: 6px 14px;
+  font-size: 13px;
+  color: #b45309;
+  background: #fef3c7;
+  border-radius: 8px;
+  text-align: center;
+
+  &.error {
+    color: #b91c1c;
+    background: #fee2e2;
+  }
 }
 </style>

frontend/src/views/inference/InferenceCreateView.vue

@@ -1,12 +1,12 @@
 <script setup lang="ts">
-import { ref, reactive, computed, onMounted } from 'vue'
+import { ref, reactive, computed, onMounted, watch } from 'vue'
 import { useRouter } from 'vue-router'
 import { ElMessage, type FormInstance, type FormRules } from 'element-plus'
 import PageCard from '@/components/PageCard.vue'
 import { getModelList, getTrainedModels } from '@/api/modules/model'
 import { getSystemInfo } from '@/api/modules/system'
 import { getComputeNodes, type ComputeNode } from '@/api/modules/compute'
-import { createCompare, preloadLocalModel, preloadTrainedModel } from '@/api/modules/compare'
+import { createCompare, loadCompare } from '@/api/modules/compare'
 import type { ModelItem, TrainedModel, GpuInfo } from '@/types'
 
 const router = useRouter()
@@ -27,6 +27,8 @@ interface SelectableModel {
   name: string
   source: 'database' | 'trained'
   model_path: string
+  compute_node_id?: string
+  compute_node_name?: string
   merged?: boolean
   merging?: boolean
   disabled?: boolean
@@ -51,6 +53,8 @@ const trainedOptions = computed<SelectableModel[]>(() =>
     name: m.name,
     source: 'trained',
     model_path: m.merged_path || m.base_model_path || '',
+    compute_node_id: m.compute_node_id,
+    compute_node_name: m.compute_node_name,
     merged: m.merged,
     merging: m.merging,
     disabled: m.merged === false,
@@ -80,7 +84,7 @@ const form = reactive({
   /** 选中的模型 key（单选） */
   model_key: '',
   /** 使用的 GPU */
-  gpu_id: 0,
+  gpu_key: '',
 })
 
 const rules: FormRules = {
@@ -90,6 +94,13 @@ const rules: FormRules = {
 
 /** 当前选中的模型对象 */
 const selectedModel = computed(() => modelMap.value[form.model_key])
+const selectedGpu = computed(() => idleGpus.value.find((g) => `${g.node_id || ''}:${g.id ?? 0}` === form.gpu_key))
+
+watch(selectedModel, (model) => {
+  if (!model?.compute_node_id) return
+  const gpu = idleGpus.value.find((item) => item.node_id === model.compute_node_id)
+  if (gpu) form.gpu_key = `${gpu.node_id || ''}:${gpu.id ?? 0}`
+})
 
 async function handleSubmit() {
   if (!formRef.value) return
@@ -101,62 +112,47 @@ async function handleSubmit() {
       return
     }
     submitting.value = true
-    startupStatus.value = '正在启动模型服务...'
+    startupStatus.value = '正在创建推理任务...'
     try {
-      // Step 1: 将模型加载到算力节点
-      const preloadPayload = {
-        model_name_or_path: m.model_path,
-        model_name: m.name,
-        template: 'qwen',
-      }
-      let preloadResult: any
-      if (m.source === 'trained') {
-        preloadResult = await preloadTrainedModel(preloadPayload)
-      } else {
-        preloadResult = await preloadLocalModel(preloadPayload)
-      }
-
-      if (preloadResult && (preloadResult as any).error) {
-        ElMessage.warning(`模型加载失败：${(preloadResult as any).error}`)
-        submitting.value = false
-        startupStatus.value = ''
+      if (!m.model_path) {
+        ElMessage.warning('当前模型未配置算力节点可访问路径，请先在模型管理中维护模型路径')
         return
       }
 
-      // Step 2: 创建推理任务记录
+      // Step 1: 创建推理任务记录
       const taskResult = await createCompare({
         name: form.name || m.name,
         description: form.description,
+        status: 'pending',
         models: [
           {
             model_id: String(m.id),
             model_name: m.name,
             model_path: m.model_path,
             source: m.source,
-            gpu_id: form.gpu_id,
+            gpu_id: selectedGpu.value?.id ?? 0,
+            node_id: selectedGpu.value?.node_id || m.compute_node_id,
+            node_name: selectedGpu.value?.node_name || m.compute_node_name,
           },
         ],
       })
       const taskId = taskResult?.id || 'unknown'
 
-      ElMessage.success('模型已启动')
-      router.push({
-        path: `/model-inference/chat/${taskId}`,
-        query: {
-          model: m.name,
-          source: m.source,
-          model_path: m.model_path,
-        },
-      })
+      // Step 2: 统一通过推理任务加载接口异步派发模型加载，状态会落到列表记录中。
+      startupStatus.value = '正在启动模型服务，首次加载可能需要数分钟...'
+      const loadResult: any = await loadCompare(taskId)
+      if (loadResult?.status === 'failed' || loadResult?.error) {
+        ElMessage.warning(`模型加载失败：${loadResult?.error || '请检查算力节点日志'}`)
+        router.push('/model-inference')
+        return
+      }
+
+      // 加载为异步派发，回到列表页可看到“启动中 → 已就绪”的状态流转
+      ElMessage.success('模型加载中，就绪后即可对话')
+      router.push('/model-inference')
     } catch (e: any) {
-      // 真实 API 失败时回退到 mock 模式（方便无算力节点的开发调试）
-      const m = selectedModel.value!
       const reason = e?.message || e?.toString() || '未知错误'
-      ElMessage.warning(`推理服务启动失败：${reason}，进入 mock 演示模式`)
-      router.push({
-        path: '/model-inference/chat/mock',
-        query: { model: m.name },
-      })
+      ElMessage.warning(`推理服务启动失败：${reason}`)
     } finally {
       submitting.value = false
       startupStatus.value = ''
@@ -181,7 +177,10 @@ async function loadData() {
     gpus.value = sys?.gpu || []
     computeNodes.value = nodes || []
     // 默认选中第一个空闲 GPU
-    if (idleGpus.value.length > 0) form.gpu_id = idleGpus.value[0].id ?? 0
+    if (idleGpus.value.length > 0) {
+      const firstGpu = idleGpus.value[0]
+      form.gpu_key = `${firstGpu.node_id || ''}:${firstGpu.id ?? 0}`
+    }
   } catch {
     // ignore
   }
@@ -229,12 +228,12 @@ onMounted(loadData)
       </el-form-item>
 
       <el-form-item label="GPU">
-        <el-select v-model="form.gpu_id" style="width: 400px">
+        <el-select v-model="form.gpu_key" style="width: 400px">
           <el-option
             v-for="g in idleGpus"
-            :key="g.id ?? 0"
-            :label="`${g.name} (GPU${g.id ?? 0}) [空闲]`"
-            :value="g.id ?? 0"
+            :key="`${g.node_id || ''}:${g.id ?? 0}`"
+            :label="`${g.node_name || g.node_code || '算力节点'} / ${g.name} (GPU${g.id ?? 0}) [空闲]`"
+            :value="`${g.node_id || ''}:${g.id ?? 0}`"
           />
         </el-select>
       </el-form-item>

frontend/src/views/inference/InferenceListView.vue

@@ -92,11 +92,9 @@ async function handleUnload(row: any) {
   loadData()
 }
 
-/** 删除（先释放算力节点再删除记录） */
+/** 删除（后端删除内部会 best-effort 释放算力节点，这里直接删记录） */
 async function handleDelete(row: any) {
   await ElMessageBox.confirm('确定要删除该推理记录吗？将先释放算力节点再删除。', '确认删除', { type: 'warning' })
-  // 先释放算力节点上的模型
-  await unloadCompare(row.id).catch(() => {})
   await deleteCompare(row.id)
   dataList.value = dataList.value.filter((item) => item.id !== row.id)
   await loadData(true)

frontend/src/views/model/MergeWeightsView.vue

@@ -18,9 +18,12 @@ const trainedModels = ref<TrainedModel[]>([])
 const currentModel = computed(() => trainedModels.value.find((m) => m.name === modelName.value))
 
 const form = reactive({
+  trained_model_id: '',
   model_name: modelName.value,
   train_method: method.value,
   base_model_path: '',
+  adapter_path: '',
+  compute_node_id: '',
 })
 
 async function loadModel() {
@@ -28,23 +31,30 @@ async function loadModel() {
     const res = await getTrainedModels()
     trainedModels.value = res?.models || []
     const target = trainedModels.value.find((m) => m.name === modelName.value)
+    form.trained_model_id = target?.id == null ? '' : String(target.id)
     form.base_model_path = target?.base_model_path || ''
+    form.adapter_path = target?.artifact_dir || target?.adapter_path || target?.merged_path || ''
+    form.compute_node_id = target?.compute_node_id || ''
   } catch {
     // ignore
   }
 }
 
 async function handleMerge() {
-  if (!form.model_name || !form.base_model_path) {
+  if (!form.model_name || !form.base_model_path || !form.adapter_path) {
     ElMessage.warning('缺少模型信息')
     return
   }
   merging.value = true
   try {
     await mergeModel({
+      trained_model_id: form.trained_model_id || form.model_name,
       model_name: form.model_name,
       train_method: form.train_method,
       base_model_path: form.base_model_path,
+      adapter_path: form.adapter_path,
+      compute_node_id: form.compute_node_id,
+      output_model_name: `${form.model_name}-merged`,
     })
     ElMessage.success('合并成功')
     router.push('/model-manage')
@@ -82,6 +92,9 @@ onMounted(loadModel)
       <el-form-item label="基座模型路径">
         <el-input v-model="form.base_model_path" placeholder="基座模型路径" />
       </el-form-item>
+      <el-form-item label="Adapter 路径">
+        <el-input v-model="form.adapter_path" placeholder="LoRA Adapter 权重目录" />
+      </el-form-item>
 
       <el-form-item>
         <el-button type="primary" :loading="merging" @click="handleMerge">

frontend/src/views/projects/ProjectListView.vue

@@ -99,7 +99,7 @@ onMounted(() => {
         <el-button type="primary" :icon="Plus" @click="showCreate = true">新建项目</el-button>
       </template>
       <template #columns>
-        <el-table-column prop="name" label="项目名" min-width="140" />
+        <el-table-column prop="name" label="项目名称" min-width="140" />
         <el-table-column prop="code" label="编码 ID" min-width="100" />
         <el-table-column prop="status" label="状态" min-width="100" />
         <el-table-column prop="description" label="描述" min-width="200" show-overflow-tooltip />
@@ -107,13 +107,13 @@ onMounted(() => {
       </template>
       <template #actions="{ row }">
         <el-button link type="primary" @click="openDetail(row.id)">详情</el-button>
-        <el-button link type="danger" @click="handleDelete(row)">删除</el-button>
+        <el-button link type="danger" @click="handleDelete(asProject(row))">删除</el-button>
       </template>
     </DataTablePage>
     <el-dialog v-model="showCreate" title="新建项目" width="520px">
       <el-form label-width="90px">
         <el-form-item label="名称" required>
-          <el-input v-model="form.name" placeholder="项目名" />
+          <el-input v-model="form.name" placeholder="项目名称" />
         </el-form-item>
         <el-form-item label="编码 ID" required>
           <el-select v-model="form.tenant_id" style="width: 100%" placeholder="选择租户编码">

frontend/src/views/system/TrainingLogView.vue

@@ -8,13 +8,14 @@ import TrainingTaskOverview from './training-log/TrainingTaskOverview.vue'
 import { usePolling } from '@/composables/usePolling'
 import '@/plugins/echarts-training-log'
 import { useModelsStore } from '@/stores/models'
-import { getFineTune, getFineTuneDiagnostics, getFineTuneLogs, type TrainingDiagnostic } from '@/api/modules/fineTune'
+import { getFineTune, getFineTuneDiagnostics, getFineTuneLogs, getFineTuneMetrics, type TrainingDiagnostic } from '@/api/modules/fineTune'
 import { getTrainingLogFiles, getTrainingLogContent } from '@/api/modules/log'
 import { getDataset } from '@/api/modules/dataset'
 import { getSystemInfo } from '@/api/modules/system'
 import { TRAIN_TYPE_MAP, TRAIN_METHOD_MAP } from '@/constants'
 import {
   buildMetricChartOption,
+  metricsFromApi,
   parseTrainingLog,
   resolveTrainingLogFile,
 } from './training-log/trainingLogModel'
@@ -42,6 +43,7 @@ const loading = ref(true)
 
 // 训练指标数据（ECharts 接收 number[]，下标即 step）
 const metricData = reactive({
+  steps: [] as number[],
   loss: [] as number[],
   gradNorm: [] as number[],
   lr: [] as number[],
@@ -62,9 +64,9 @@ const gpuExpanded = ref(false)
 let refreshInFlight = false
 
 /** 三个曲线的 ECharts 配置（响应式，数据变化自动重绘） */
-const lossChartOption = computed(() => buildMetricChartOption('Loss', metricData.loss, '#4f46e5'))
-const gradChartOption = computed(() => buildMetricChartOption('Grad Norm', metricData.gradNorm, '#3b82f6'))
-const lrChartOption = computed(() => buildMetricChartOption('Learning Rate', metricData.lr, '#14b8a6', true))
+const lossChartOption = computed(() => buildMetricChartOption('Loss', metricData.loss, metricData.steps, '#4f46e5'))
+const gradChartOption = computed(() => buildMetricChartOption('Grad Norm', metricData.gradNorm, metricData.steps, '#3b82f6'))
+const lrChartOption = computed(() => buildMetricChartOption('Learning Rate', metricData.lr, metricData.steps, '#14b8a6', true))
 const baseModelName = computed(() => task.value?.base_model != null
   ? modelsStore.getModelName(task.value.base_model)
   : '未配置')
@@ -77,10 +79,10 @@ const trainingMethodName = computed(() => task.value?.train_method
 const taskGpuLabel = computed(() => task.value?.gpus?.length
   ? task.value.gpus.map((gpuId) => `GPU ${gpuId}`).join('、')
   : '未配置')
-const latestLoss = computed(() => metricData.loss[metricData.loss.length - 1])
-const latestGradNorm = computed(() => metricData.gradNorm[metricData.gradNorm.length - 1])
-const latestLearningRate = computed(() => metricData.lr[metricData.lr.length - 1])
-const latestEpoch = computed(() => metricData.epoch[metricData.epoch.length - 1])
+const latestLoss = computed(() => lastFinite(metricData.loss))
+const latestGradNorm = computed(() => lastFinite(metricData.gradNorm))
+const latestLearningRate = computed(() => lastFinite(metricData.lr))
+const latestEpoch = computed(() => lastFinite(metricData.epoch))
 const logLineCount = computed(() => logContent.value ? logContent.value.split(/\r?\n/).length : 0)
 const taskGpuItems = computed<TaskGpuItem[]>(() => (task.value?.gpus ?? []).map((gpuId) => {
   const index = Number(gpuId)
@@ -123,7 +125,7 @@ const gpuRefreshState = computed(() => {
   })
   return gpuLoadError.value
     ? `更新失败 · 最后更新 ${updateTime}`
-    : `${updateTime} 更新 · 每 5 秒刷新`
+    : `${updateTime} 更新 · 每 3 秒刷新`
 })
 
 function formatMetric(value?: number, scientific = false) {
@@ -131,6 +133,13 @@ function formatMetric(value?: number, scientific = false) {
   return scientific ? value.toExponential(2) : value.toFixed(4).replace(/0+$/, '').replace(/\.$/, '')
 }
 
+function lastFinite(values: number[]) {
+  for (let index = values.length - 1; index >= 0; index -= 1) {
+    if (Number.isFinite(values[index])) return values[index]
+  }
+  return undefined
+}
+
 function safePercent(value?: number) {
   return Math.round(Math.min(100, Math.max(0, Number(value || 0))))
 }
@@ -216,6 +225,7 @@ const isLoraMethod = computed(() =>
 function applyLogContent(content: string) {
   const parsed = parseTrainingLog(content)
   logContent.value = content
+  metricData.steps = parsed.metrics.steps
   metricData.loss = parsed.metrics.loss
   metricData.gradNorm = parsed.metrics.gradNorm
   metricData.lr = parsed.metrics.lr
@@ -223,6 +233,26 @@ function applyLogContent(content: string) {
   Object.assign(summary, parsed.summary)
 }
 
+function applyMetricData(metrics = { steps: [] as number[], loss: [] as number[], gradNorm: [] as number[], lr: [] as number[], epoch: [] as number[] }) {
+  metricData.steps = metrics.steps
+  metricData.loss = metrics.loss
+  metricData.gradNorm = metrics.gradNorm
+  metricData.lr = metrics.lr
+  metricData.epoch = metrics.epoch
+}
+
+async function loadMetrics(currentTask: FineTuneTask) {
+  try {
+    const points = await getFineTuneMetrics(currentTask.id)
+    const parsed = metricsFromApi(points || [])
+    if (parsed.loss.length || parsed.gradNorm.length || parsed.lr.length) {
+      applyMetricData(parsed)
+    }
+  } catch {
+    // 日志解析结果会作为兜底曲线数据。
+  }
+}
+
 async function loadLog(currentTask: FineTuneTask) {
   try {
     const runtime = await getFineTuneLogs(currentTask.id, { tail_lines: 800 })
@@ -277,6 +307,7 @@ async function refreshAll() {
       ? loadDataset(currentTask.train_dataset_id)
       : Promise.resolve()
     await Promise.all([datasetPromise, loadLog(currentTask), loadGpuStatus(), loadDiagnostics(currentTask)])
+    await loadMetrics(currentTask)
   } finally {
     loading.value = false
     refreshInFlight = false
@@ -523,7 +554,7 @@ onMounted(async () => {
 
     <!-- 训练曲线 -->
     <PageCard class="metrics-panel" title="训练曲线" subtitle="持续监控模型收敛情况与学习率变化">
-      <template #extra><span class="refresh-state">每 5 秒刷新</span></template>
+      <template #extra><span class="refresh-state">每 3 秒刷新</span></template>
       <div class="chart-list" aria-label="训练指标曲线">
         <section class="chart-section">
           <div class="chart-section-header">
@@ -560,7 +591,7 @@ onMounted(async () => {
 
     <!-- 原始日志 -->
     <PageCard class="log-card" title="训练日志" subtitle="查看训练任务的原始运行输出">
-      <template #extra><span class="log-meta">{{ logLineCount }} 行 · 每 5 秒刷新</span></template>
+      <template #extra><span class="log-meta">{{ logLineCount }} 行 · 每 3 秒刷新</span></template>
       <pre class="log-pre">{{ logContent || '暂无日志' }}</pre>
     </PageCard>
     </template>

frontend/src/views/system/training-log/trainingLogModel.ts

@@ -1,7 +1,9 @@
 import type { EChartsOption } from 'echarts'
 import type { FineTuneTask, TrainingLogFile } from '@/types'
+import type { FineTuneMetricPoint } from '@/api/modules/fineTune'
 
 export interface TrainingMetricData {
+  steps: number[]
   loss: number[]
   gradNorm: number[]
   lr: number[]
@@ -26,7 +28,7 @@ function escapeRegExp(value: string) {
 }
 
 function extractNumber(source: string, key: string) {
-  const match = source.match(new RegExp(`['"]?${escapeRegExp(key)}['"]?\\s*:\\s*(${NUMBER_SOURCE})`, 'i'))
+  const match = source.match(new RegExp(`['"]?${escapeRegExp(key)}['"]?\\s*(?:=|:)\\s*(${NUMBER_SOURCE})`, 'i'))
   return match ? Number(match[1]) : undefined
 }
 
@@ -57,24 +59,44 @@ export function resolveTrainingLogFile(
 
 /** 解析日志中的逐步指标。字段顺序和常见数值格式均不受限制。 */
 export function parseTrainingMetrics(text: string): TrainingMetricData {
-  const metrics: TrainingMetricData = { loss: [], gradNorm: [], lr: [], epoch: [] }
-  const blocks = text.match(/\{[^{}\r\n]*\}/g) || []
+  const metrics: TrainingMetricData = { steps: [], loss: [], gradNorm: [], lr: [], epoch: [] }
+  const candidates = text
+    .split(/\r?\n/)
+    .flatMap((line) => {
+      const blocks = line.match(/\{[^{}\r\n]*\}/g)
+      return blocks?.length ? blocks.map((block) => `${line} ${block}`) : [line]
+    })
 
-  for (const block of blocks) {
-    const loss = extractNumber(block, 'loss')
-    const gradNorm = extractNumber(block, 'grad_norm')
-    const learningRate = extractNumber(block, 'learning_rate')
-    const epoch = extractNumber(block, 'epoch')
-    if (loss == null || gradNorm == null || learningRate == null) continue
-    metrics.loss.push(loss)
-    metrics.gradNorm.push(gradNorm)
-    metrics.lr.push(learningRate)
-    if (epoch != null) metrics.epoch.push(epoch)
+  for (const [index, line] of candidates.entries()) {
+    const loss = extractNumber(line, 'loss')
+    const gradNorm = extractNumber(line, 'grad_norm')
+    const learningRate = extractNumber(line, 'learning_rate')
+    const epoch = extractNumber(line, 'epoch')
+    if (loss == null && gradNorm == null && learningRate == null) continue
+    metrics.steps.push(extractNumber(line, 'step') ?? metrics.steps.length + index + 1)
+    metrics.loss.push(loss ?? Number.NaN)
+    metrics.gradNorm.push(gradNorm ?? Number.NaN)
+    metrics.lr.push(learningRate ?? Number.NaN)
+    metrics.epoch.push(epoch ?? Number.NaN)
   }
 
   return metrics
 }
 
+export function metricsFromApi(points: FineTuneMetricPoint[]): TrainingMetricData {
+  const metrics: TrainingMetricData = { steps: [], loss: [], gradNorm: [], lr: [], epoch: [] }
+  for (const [index, point] of points.entries()) {
+    const hasMetric = point.loss != null || point.grad_norm != null || point.learning_rate != null
+    if (!hasMetric) continue
+    metrics.steps.push(Number(point.step || index + 1))
+    metrics.loss.push(point.loss == null ? Number.NaN : Number(point.loss))
+    metrics.gradNorm.push(point.grad_norm == null ? Number.NaN : Number(point.grad_norm))
+    metrics.lr.push(point.learning_rate == null ? Number.NaN : Number(point.learning_rate))
+    metrics.epoch.push(point.epoch == null ? Number.NaN : Number(point.epoch))
+  }
+  return metrics
+}
+
 /** 每次都返回新对象，日志截断或切换时不会残留上一轮汇总。 */
 export function parseTrainingSummary(text: string): TrainingSummary {
   const emptySummary: TrainingSummary = { epoch: '', trainLoss: '', runtime: '' }
@@ -102,11 +124,23 @@ export function parseTrainingLog(text: string): ParsedTrainingLog {
 export function buildMetricChartOption(
   label: string,
   data: number[],
+  steps: number[],
   color: string,
   logScale = false,
 ): EChartsOption {
+  const visibleData = data.map((value) => (Number.isFinite(value) ? value : null))
   return {
     grid: { top: 24, right: 20, bottom: 56, left: 56 },
+    graphic: visibleData.some((value) => value != null)
+      ? []
+      : [
+          {
+            type: 'text',
+            left: 'center',
+            top: 'middle',
+            style: { text: '暂无训练指标数据', fill: '#94a3b8', fontSize: 13 },
+          },
+        ],
     tooltip: {
       trigger: 'axis',
       axisPointer: { type: 'cross' },
@@ -116,6 +150,7 @@ export function buildMetricChartOption(
     },
     xAxis: {
       type: 'category',
+      data: steps.map((step, index) => (Number.isFinite(step) ? String(step) : String(index + 1))),
       boundaryGap: false,
       name: 'Step',
       nameTextStyle: { color: '#94a3b8', fontSize: 11 },
@@ -142,7 +177,7 @@ export function buildMetricChartOption(
       {
         name: label,
         type: 'line',
-        data,
+        data: visibleData,
         smooth: true,
         symbol: 'none',
         lineStyle: { width: 2, color },

frontend/src/views/tenants/TenantListView.vue

@@ -30,7 +30,7 @@ function formatQuota(quota: Record<string, unknown> | undefined | null) {
   if (q.gpu > 0) parts.push(`GPU ${q.gpu}`)
   if (q.storage > 0) parts.push(`存储 ${q.storage}GB`)
   if (q.maxProjects > 0) parts.push(`项目 ${q.maxProjects}`)
-  return parts.length ? parts.join(' | ') : '—'
+  return parts.length ? parts.join(' | ') : '-'
 }
 
 async function load() {
@@ -52,7 +52,7 @@ function asTenant(row: unknown): Tenant {
 
 function quotaText(row: unknown): string {
   const quota = asTenant(row).quota || {}
-  return Object.keys(quota).length ? JSON.stringify(quota) : '—'
+  return Object.keys(quota).length ? JSON.stringify(quota) : '-'
 }
 
 async function submitCreate() {
@@ -118,13 +118,13 @@ onMounted(load)
       </template>
       <template #columns>
         <el-table-column prop="name" label="租户名称" min-width="140" />
-        <el-table-column prop="code" label="用户ID" min-width="100" />
+        <el-table-column prop="code" label="租户 ID" min-width="100" />
         <el-table-column prop="status" label="状态" min-width="100" />
         <el-table-column prop="create_time" label="创建时间" min-width="180" />
       </template>
       <template #actions="{ row }">
         <el-button link type="primary" @click="openDetail(row.id)">详情</el-button>
-        <el-button link type="danger" @click="handleDelete(row)">删除</el-button>
+        <el-button link type="danger" @click="handleDelete(asTenant(row))">删除</el-button>
       </template>
     </DataTablePage>