Upload merged Qwen3-4B-Instruct-2507 model (auto-generated README)
Browse files- README.md +2 -16
- model-00001-of-00002.safetensors +1 -1
- model-00002-of-00002.safetensors +1 -1
README.md
CHANGED
|
@@ -1,14 +1,7 @@
|
|
| 1 |
---
|
| 2 |
base_model: Qwen/Qwen3-4B-Instruct-2507
|
| 3 |
datasets:
|
| 4 |
-
- u-10bei/sft_alfworld_trajectory_dataset
|
| 5 |
-
- u-10bei/sft_alfworld_trajectory_dataset_v2
|
| 6 |
-
- u-10bei/sft_alfworld_trajectory_dataset_v3
|
| 7 |
-
- u-10bei/sft_alfworld_trajectory_dataset_v4
|
| 8 |
- u-10bei/sft_alfworld_trajectory_dataset_v5
|
| 9 |
-
- u-10bei/dbbench_sft_dataset_react
|
| 10 |
-
- u-10bei/dbbench_sft_dataset_react_v2
|
| 11 |
-
- u-10bei/dbbench_sft_dataset_react_v3
|
| 12 |
- u-10bei/dbbench_sft_dataset_react_v4
|
| 13 |
language:
|
| 14 |
- en
|
|
@@ -45,7 +38,7 @@ tool use, and recovery from errors.
|
|
| 45 |
- Base model: Qwen/Qwen3-4B-Instruct-2507
|
| 46 |
- Method: LoRA (full precision base)
|
| 47 |
- Max sequence length: 2048
|
| 48 |
-
- Epochs:
|
| 49 |
- Learning rate: 2e-06
|
| 50 |
- LoRA: r=64, alpha=128
|
| 51 |
|
|
@@ -70,15 +63,8 @@ model = PeftModel.from_pretrained(model, adapter)
|
|
| 70 |
|
| 71 |
## Sources & Terms (IMPORTANT)
|
| 72 |
|
| 73 |
-
Training data:
|
| 74 |
-
- u-10bei/sft_alfworld_trajectory_dataset
|
| 75 |
-
- u-10bei/sft_alfworld_trajectory_dataset_v2
|
| 76 |
-
- u-10bei/sft_alfworld_trajectory_dataset_v3
|
| 77 |
-
- u-10bei/sft_alfworld_trajectory_dataset_v4
|
| 78 |
- u-10bei/sft_alfworld_trajectory_dataset_v5
|
| 79 |
-
- u-10bei/dbbench_sft_dataset_react
|
| 80 |
-
- u-10bei/dbbench_sft_dataset_react_v2
|
| 81 |
-
- u-10bei/dbbench_sft_dataset_react_v3
|
| 82 |
- u-10bei/dbbench_sft_dataset_react_v4
|
| 83 |
|
| 84 |
Dataset License: MIT License. This dataset is used and distributed under the terms of the MIT License.
|
|
|
|
| 1 |
---
|
| 2 |
base_model: Qwen/Qwen3-4B-Instruct-2507
|
| 3 |
datasets:
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
- u-10bei/sft_alfworld_trajectory_dataset_v5
|
|
|
|
|
|
|
|
|
|
| 5 |
- u-10bei/dbbench_sft_dataset_react_v4
|
| 6 |
language:
|
| 7 |
- en
|
|
|
|
| 38 |
- Base model: Qwen/Qwen3-4B-Instruct-2507
|
| 39 |
- Method: LoRA (full precision base)
|
| 40 |
- Max sequence length: 2048
|
| 41 |
+
- Epochs: 2
|
| 42 |
- Learning rate: 2e-06
|
| 43 |
- LoRA: r=64, alpha=128
|
| 44 |
|
|
|
|
| 63 |
|
| 64 |
## Sources & Terms (IMPORTANT)
|
| 65 |
|
| 66 |
+
Training data:
|
|
|
|
|
|
|
|
|
|
|
|
|
| 67 |
- u-10bei/sft_alfworld_trajectory_dataset_v5
|
|
|
|
|
|
|
|
|
|
| 68 |
- u-10bei/dbbench_sft_dataset_react_v4
|
| 69 |
|
| 70 |
Dataset License: MIT License. This dataset is used and distributed under the terms of the MIT License.
|
model-00001-of-00002.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4967215360
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e104db3f42afba923817940d1f69cc420c786d6fbcdd7020bcc48dbc27ebdeda
|
| 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:
|
| 3 |
size 3077766632
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cf4a2cfdab371cbd20a6f44c4d57ac183405d06b60b01f88ec7731e63e252adf
|
| 3 |
size 3077766632
|