Hugging Face's logo

unsloth
/
Qwen3-8B-Base-unsloth-bnb-4bit

Text Generation
Transformers
Safetensors
qwen3
unsloth
text-generation-inference
4-bit precision
bitsandbytes
Model card Files
xet
Community
1
danielhanchen commited on
Commit
c49b94d
·
verified ·
1 Parent(s): 768be47

Upload folder using huggingface_hub

Browse files
Files changed (5) hide show
  1. README.md +25 -6
  2. config.json +42 -2
  3. generation_config.json +2 -2
  4. special_tokens_map.json +1 -1
  5. tokenizer_config.json +1 -1
README.md CHANGED
@@ -6,6 +6,23 @@ base_model:
6
  license: apache-2.0
7
  library_name: transformers
8
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9
  # Qwen3-8B-Base
10
 
11
  ## Qwen3 Highlights
@@ -49,11 +66,13 @@ Detailed evaluation results are reported in this [📑 blog](https://qwenlm.gith
49
  If you find our work helpful, feel free to give us a cite.
50
 
51
  ```
52
- @misc{qwen3,
53
- title = {Qwen3},
54
- url = {https://qwenlm.github.io/blog/qwen3/},
55
- author = {Qwen Team},
56
- month = {April},
57
- year = {2025}
 
 
58
  }
59
  ```
 
6
  license: apache-2.0
7
  library_name: transformers
8
  ---
9
+ <div>
10
+ <p style="margin-top: 0;margin-bottom: 0;">
11
+ <em><a href="https://docs.unsloth.ai/basics/unsloth-dynamic-v2.0-gguf">Unsloth Dynamic 2.0</a> achieves superior accuracy & outperforms other leading quants.</em>
12
+ </p>
13
+ <div style="display: flex; gap: 5px; align-items: center; ">
14
+ <a href="https://github.com/unslothai/unsloth/">
15
+ <img src="https://github.com/unslothai/unsloth/raw/main/images/unsloth%20new%20logo.png" width="133">
16
+ </a>
17
+ <a href="https://discord.gg/unsloth">
18
+ <img src="https://github.com/unslothai/unsloth/raw/main/images/Discord%20button.png" width="173">
19
+ </a>
20
+ <a href="https://docs.unsloth.ai/">
21
+ <img src="https://raw.githubusercontent.com/unslothai/unsloth/refs/heads/main/images/documentation%20green%20button.png" width="143">
22
+ </a>
23
+ </div>
24
+ </div>
25
+
26
  # Qwen3-8B-Base
27
 
28
  ## Qwen3 Highlights
 
66
  If you find our work helpful, feel free to give us a cite.
67
 
68
  ```
69
+ @misc{qwen3technicalreport,
70
+ title={Qwen3 Technical Report},
71
+ author={Qwen Team},
72
+ year={2025},
73
+ eprint={2505.09388},
74
+ archivePrefix={arXiv},
75
+ primaryClass={cs.CL},
76
+ url={https://arxiv.org/abs/2505.09388},
77
  }
78
  ```
config.json CHANGED
@@ -4,12 +4,50 @@
4
  ],
5
  "attention_bias": false,
6
  "attention_dropout": 0.0,
7
- "eos_token_id": 151645,
8
  "head_dim": 128,
9
  "hidden_act": "silu",
10
  "hidden_size": 4096,
11
  "initializer_range": 0.02,
12
  "intermediate_size": 12288,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
13
  "max_position_embeddings": 32768,
14
  "max_window_layers": 36,
15
  "model_type": "qwen3",
@@ -27,6 +65,8 @@
27
  "llm_int8_enable_fp32_cpu_offload": false,
28
  "llm_int8_has_fp16_weight": false,
29
  "llm_int8_skip_modules": [
 
 
30
  "lm_head",
31
  "multi_modal_projector",
32
  "merger",
@@ -46,7 +86,7 @@
46
  "sliding_window": null,
47
  "tie_word_embeddings": false,
48
  "torch_dtype": "bfloat16",
49
- "transformers_version": "4.51.3",
50
  "unsloth_fixed": true,
51
  "use_cache": true,
52
  "use_sliding_window": false,
 
4
  ],
5
  "attention_bias": false,
6
  "attention_dropout": 0.0,
7
+ "eos_token_id": 151643,
8
  "head_dim": 128,
9
  "hidden_act": "silu",
10
  "hidden_size": 4096,
11
  "initializer_range": 0.02,
12
  "intermediate_size": 12288,
13
+ "layer_types": [
14
+ "full_attention",
15
+ "full_attention",
16
+ "full_attention",
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention",
45
+ "full_attention",
46
+ "full_attention",
47
+ "full_attention",
48
+ "full_attention",
49
+ "full_attention"
50
+ ],
51
  "max_position_embeddings": 32768,
52
  "max_window_layers": 36,
53
  "model_type": "qwen3",
 
65
  "llm_int8_enable_fp32_cpu_offload": false,
66
  "llm_int8_has_fp16_weight": false,
67
  "llm_int8_skip_modules": [
68
+ "embed_tokens",
69
+ "embedding",
70
  "lm_head",
71
  "multi_modal_projector",
72
  "merger",
 
86
  "sliding_window": null,
87
  "tie_word_embeddings": false,
88
  "torch_dtype": "bfloat16",
89
+ "transformers_version": "4.53.2",
90
  "unsloth_fixed": true,
91
  "use_cache": true,
92
  "use_sliding_window": false,
generation_config.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "bos_token_id": 151643,
3
- "eos_token_id": 151645,
4
  "max_length": 32768,
5
  "max_new_tokens": 2048,
6
  "pad_token_id": 151654,
7
- "transformers_version": "4.51.3"
8
  }
 
1
  {
2
  "bos_token_id": 151643,
3
+ "eos_token_id": 151643,
4
  "max_length": 32768,
5
  "max_new_tokens": 2048,
6
  "pad_token_id": 151654,
7
+ "transformers_version": "4.53.2"
8
  }
special_tokens_map.json CHANGED
@@ -15,7 +15,7 @@
15
  "<|video_pad|>"
16
  ],
17
  "eos_token": {
18
- "content": "<|im_end|>",
19
  "lstrip": false,
20
  "normalized": false,
21
  "rstrip": false,
 
15
  "<|video_pad|>"
16
  ],
17
  "eos_token": {
18
+ "content": "<|endoftext|>",
19
  "lstrip": false,
20
  "normalized": false,
21
  "rstrip": false,
tokenizer_config.json CHANGED
@@ -228,7 +228,7 @@
228
  ],
229
  "bos_token": null,
230
  "clean_up_tokenization_spaces": false,
231
- "eos_token": "<|im_end|>",
232
  "errors": "replace",
233
  "extra_special_tokens": {},
234
  "model_max_length": 32768,
 
228
  ],
229
  "bos_token": null,
230
  "clean_up_tokenization_spaces": false,
231
+ "eos_token": "<|endoftext|>",
232
  "errors": "replace",
233
  "extra_special_tokens": {},
234
  "model_max_length": 32768,