AF0815 commited on
Commit
f855213
·
verified ·
1 Parent(s): 843d2b5

Upload merged Qwen3-4B-Instruct-2507 model (auto-generated README)

Browse files
README.md CHANGED
@@ -65,9 +65,9 @@ DB category weights used during training-data preparation:
65
  - Base model: Qwen/Qwen3-4B-Instruct-2507
66
  - Method: LoRA (full precision base)
67
  - Max sequence length: 2048
68
- - Epochs: 1.2
69
  - Learning rate: 3e-06
70
- - LoRA: r=32, alpha=64, dropout=0.0
71
  - Per-device train batch size: 2
72
  - Gradient accumulation: 4
73
 
 
65
  - Base model: Qwen/Qwen3-4B-Instruct-2507
66
  - Method: LoRA (full precision base)
67
  - Max sequence length: 2048
68
+ - Epochs: 2.0
69
  - Learning rate: 3e-06
70
+ - LoRA: r=32, alpha=64, dropout=0.05
71
  - Per-device train batch size: 2
72
  - Gradient accumulation: 4
73
 
model-00001-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:138c634ad430d7e57b870bbf5595d5c5cd81d8b4a0d0000848336893219da68d
3
  size 4967215360
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:05c08cd2358f4731239d0852327bc3557cfd3dfa2a35642a3f1d822329035722
3
  size 4967215360
model-00002-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:35dd175b4f465cc7477e750b412dbf66c06fdf110ba534e33beb8680e535517b
3
  size 3077766632
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3540058ae454cb87e01380ab5910784fc676ba5349e85b232446df661bffb673
3
  size 3077766632