[LAB 01 · 固定表示：标签预算与读出](/course/notebooks/4dd8605ed9904820919ea9e13d9e7d7c.html)

# 本地保存输出（非 Deepnote 云端运行）
[独立阅读](/course/notebook-blocks/387907024b124f8c8dfc828f39acf1f9.html)
按问题继续

固定CNN与缓存特征、标签预算、kNN/linear和独立评价

**本地保存输出（非 Deepnote 云端运行）**

```
{
  "identity": "v17 independent frozen readout",
  "fresh_process": true,
  "local_run": true,
  "cloud_run": false,
  "source_model_unchanged": true,
  "source_target_sample_overlap": false,
  "source_target_writer_overlap": "unknown",
  "heldout_used_for_tuning": false,
  "clip_loaded": false,
  "environment": {
    "python": "3.12.14",
    "torch": "2.14.0+cpu",
    "numpy": "2.3.5",
    "sklearn": "1.8.0"
  },
  "checkpoint_sha256": "ba8b738cc67d4ec3cd83d75a5dbe8a70ec548377335ba0abaaa7c23cd14eeba3",
  "target_sha256": "ca069f616b511b5aa9d3e9ab88b6fd12741a5b9c1b66a8befe467e44a796ddd1"
}

```