dulimov commited on
Commit
4f4e8c2
·
verified ·
1 Parent(s): f88ac12

Upload folder using huggingface_hub

Browse files
Files changed (31) hide show
  1. Qwen3-0.6B-rk3588-w8a8-opt-0-hybrid-ratio-0.0.rkllm +2 -2
  2. Qwen3-0.6B-rk3588-w8a8-opt-0-hybrid-ratio-0.5.rkllm +2 -2
  3. Qwen3-0.6B-rk3588-w8a8-opt-0-hybrid-ratio-1.0.rkllm +2 -2
  4. Qwen3-0.6B-rk3588-w8a8-opt-1-hybrid-ratio-0.0.rkllm +2 -2
  5. Qwen3-0.6B-rk3588-w8a8-opt-1-hybrid-ratio-0.5.rkllm +2 -2
  6. Qwen3-0.6B-rk3588-w8a8-opt-1-hybrid-ratio-1.0.rkllm +2 -2
  7. Qwen3-0.6B-rk3588-w8a8_g128-opt-0-hybrid-ratio-0.0.rkllm +2 -2
  8. Qwen3-0.6B-rk3588-w8a8_g128-opt-0-hybrid-ratio-0.5.rkllm +2 -2
  9. Qwen3-0.6B-rk3588-w8a8_g128-opt-0-hybrid-ratio-1.0.rkllm +2 -2
  10. Qwen3-0.6B-rk3588-w8a8_g128-opt-1-hybrid-ratio-0.0.rkllm +2 -2
  11. Qwen3-0.6B-rk3588-w8a8_g128-opt-1-hybrid-ratio-0.5.rkllm +2 -2
  12. Qwen3-0.6B-rk3588-w8a8_g128-opt-1-hybrid-ratio-1.0.rkllm +2 -2
  13. Qwen3-0.6B-rk3588-w8a8_g256-opt-0-hybrid-ratio-0.0.rkllm +2 -2
  14. Qwen3-0.6B-rk3588-w8a8_g256-opt-0-hybrid-ratio-0.5.rkllm +2 -2
  15. Qwen3-0.6B-rk3588-w8a8_g256-opt-0-hybrid-ratio-1.0.rkllm +2 -2
  16. Qwen3-0.6B-rk3588-w8a8_g256-opt-1-hybrid-ratio-0.0.rkllm +2 -2
  17. Qwen3-0.6B-rk3588-w8a8_g256-opt-1-hybrid-ratio-0.5.rkllm +2 -2
  18. Qwen3-0.6B-rk3588-w8a8_g256-opt-1-hybrid-ratio-1.0.rkllm +2 -2
  19. Qwen3-0.6B-rk3588-w8a8_g512-opt-0-hybrid-ratio-0.0.rkllm +2 -2
  20. Qwen3-0.6B-rk3588-w8a8_g512-opt-0-hybrid-ratio-0.5.rkllm +2 -2
  21. Qwen3-0.6B-rk3588-w8a8_g512-opt-0-hybrid-ratio-1.0.rkllm +2 -2
  22. Qwen3-0.6B-rk3588-w8a8_g512-opt-1-hybrid-ratio-0.0.rkllm +2 -2
  23. Qwen3-0.6B-rk3588-w8a8_g512-opt-1-hybrid-ratio-0.5.rkllm +2 -2
  24. Qwen3-0.6B-rk3588-w8a8_g512-opt-1-hybrid-ratio-1.0.rkllm +2 -2
  25. README.md +11 -22
  26. added_tokens.json +28 -0
  27. chat_template.jinja +97 -0
  28. config.json +4 -3
  29. generation_config.json +13 -12
  30. special_tokens_map.json +31 -0
  31. tokenizer_config.json +7 -5
Qwen3-0.6B-rk3588-w8a8-opt-0-hybrid-ratio-0.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:06a2c870f8207d18e314c7f67d0f232cd0bb936fdbfd82c422dfce018dd35be6
3
- size 932189294
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33eaa663d464507081c1b8cb8b076b0733cf4e1df313a85f4f6ef480a93f8179
3
+ size 932189806
Qwen3-0.6B-rk3588-w8a8-opt-0-hybrid-ratio-0.5.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:06a2c870f8207d18e314c7f67d0f232cd0bb936fdbfd82c422dfce018dd35be6
3
- size 932189294
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33eaa663d464507081c1b8cb8b076b0733cf4e1df313a85f4f6ef480a93f8179
3
+ size 932189806
Qwen3-0.6B-rk3588-w8a8-opt-0-hybrid-ratio-1.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:06a2c870f8207d18e314c7f67d0f232cd0bb936fdbfd82c422dfce018dd35be6
3
- size 932189294
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33eaa663d464507081c1b8cb8b076b0733cf4e1df313a85f4f6ef480a93f8179
3
+ size 932189806
Qwen3-0.6B-rk3588-w8a8-opt-1-hybrid-ratio-0.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:06a2c870f8207d18e314c7f67d0f232cd0bb936fdbfd82c422dfce018dd35be6
3
- size 932189294
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33eaa663d464507081c1b8cb8b076b0733cf4e1df313a85f4f6ef480a93f8179
3
+ size 932189806
Qwen3-0.6B-rk3588-w8a8-opt-1-hybrid-ratio-0.5.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:06a2c870f8207d18e314c7f67d0f232cd0bb936fdbfd82c422dfce018dd35be6
3
- size 932189294
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33eaa663d464507081c1b8cb8b076b0733cf4e1df313a85f4f6ef480a93f8179
3
+ size 932189806
Qwen3-0.6B-rk3588-w8a8-opt-1-hybrid-ratio-1.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:06a2c870f8207d18e314c7f67d0f232cd0bb936fdbfd82c422dfce018dd35be6
3
- size 932189294
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33eaa663d464507081c1b8cb8b076b0733cf4e1df313a85f4f6ef480a93f8179
3
+ size 932189806
Qwen3-0.6B-rk3588-w8a8_g128-opt-0-hybrid-ratio-0.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5aa9ef530dadfc9d7d244b6e6e2e9102a199e560f3c94ecf126f9bedd55b9350
3
- size 1005830694
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab0872969b9f6db51b16b4436432a1f1562d5f9e45399ff7a6e0b1f35ec4bded
3
+ size 1005831206
Qwen3-0.6B-rk3588-w8a8_g128-opt-0-hybrid-ratio-0.5.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5aa9ef530dadfc9d7d244b6e6e2e9102a199e560f3c94ecf126f9bedd55b9350
3
- size 1005830694
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab0872969b9f6db51b16b4436432a1f1562d5f9e45399ff7a6e0b1f35ec4bded
3
+ size 1005831206
Qwen3-0.6B-rk3588-w8a8_g128-opt-0-hybrid-ratio-1.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5aa9ef530dadfc9d7d244b6e6e2e9102a199e560f3c94ecf126f9bedd55b9350
3
- size 1005830694
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab0872969b9f6db51b16b4436432a1f1562d5f9e45399ff7a6e0b1f35ec4bded
3
+ size 1005831206
Qwen3-0.6B-rk3588-w8a8_g128-opt-1-hybrid-ratio-0.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:aeda4c20fff12eadb8865af804d0dc64508d19e63cc4cff1dafc8b7a64797bf0
3
- size 996371654
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:922091b53a3698445df53ddebc19d7935aba4e303f8ea3027bf40dd41bfc0d41
3
+ size 996372166
Qwen3-0.6B-rk3588-w8a8_g128-opt-1-hybrid-ratio-0.5.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:aeda4c20fff12eadb8865af804d0dc64508d19e63cc4cff1dafc8b7a64797bf0
3
- size 996371654
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:922091b53a3698445df53ddebc19d7935aba4e303f8ea3027bf40dd41bfc0d41
3
+ size 996372166
Qwen3-0.6B-rk3588-w8a8_g128-opt-1-hybrid-ratio-1.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:aeda4c20fff12eadb8865af804d0dc64508d19e63cc4cff1dafc8b7a64797bf0
3
- size 996371654
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:922091b53a3698445df53ddebc19d7935aba4e303f8ea3027bf40dd41bfc0d41
3
+ size 996372166
Qwen3-0.6B-rk3588-w8a8_g256-opt-0-hybrid-ratio-0.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b4cdf5f13e52ed00ba59952f42f0c33e9390f282ce0e9b81969e9fb73c4f2cb1
3
- size 969551974
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a624af4291ab59aa458b9ab9d03dd71b07034f55cc0aa313e8776b8ceb672922
3
+ size 969552486
Qwen3-0.6B-rk3588-w8a8_g256-opt-0-hybrid-ratio-0.5.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b4cdf5f13e52ed00ba59952f42f0c33e9390f282ce0e9b81969e9fb73c4f2cb1
3
- size 969551974
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a624af4291ab59aa458b9ab9d03dd71b07034f55cc0aa313e8776b8ceb672922
3
+ size 969552486
Qwen3-0.6B-rk3588-w8a8_g256-opt-0-hybrid-ratio-1.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b4cdf5f13e52ed00ba59952f42f0c33e9390f282ce0e9b81969e9fb73c4f2cb1
3
- size 969551974
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a624af4291ab59aa458b9ab9d03dd71b07034f55cc0aa313e8776b8ceb672922
3
+ size 969552486
Qwen3-0.6B-rk3588-w8a8_g256-opt-1-hybrid-ratio-0.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c867729edbe1249fc2acfc2e3edfcaa2b274cc7b83d61c78b6f73c0006aa2ccc
3
- size 965459654
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26b317c545c856b8604bda828c576653d6a326073188043f59043a25ccc39de0
3
+ size 965460166
Qwen3-0.6B-rk3588-w8a8_g256-opt-1-hybrid-ratio-0.5.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c867729edbe1249fc2acfc2e3edfcaa2b274cc7b83d61c78b6f73c0006aa2ccc
3
- size 965459654
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26b317c545c856b8604bda828c576653d6a326073188043f59043a25ccc39de0
3
+ size 965460166
Qwen3-0.6B-rk3588-w8a8_g256-opt-1-hybrid-ratio-1.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c867729edbe1249fc2acfc2e3edfcaa2b274cc7b83d61c78b6f73c0006aa2ccc
3
- size 965459654
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26b317c545c856b8604bda828c576653d6a326073188043f59043a25ccc39de0
3
+ size 965460166
Qwen3-0.6B-rk3588-w8a8_g512-opt-0-hybrid-ratio-0.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0a18491f63efe98c4e4aba43b36034e4218fbe02ccd6247c9363e03ede381850
3
- size 951412614
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a7a1e7b09af10a4e60d0cff4a2e275f9b2b95df54b413fcae1cab54981c4c7e
3
+ size 951413126
Qwen3-0.6B-rk3588-w8a8_g512-opt-0-hybrid-ratio-0.5.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0a18491f63efe98c4e4aba43b36034e4218fbe02ccd6247c9363e03ede381850
3
- size 951412614
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a7a1e7b09af10a4e60d0cff4a2e275f9b2b95df54b413fcae1cab54981c4c7e
3
+ size 951413126
Qwen3-0.6B-rk3588-w8a8_g512-opt-0-hybrid-ratio-1.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0a18491f63efe98c4e4aba43b36034e4218fbe02ccd6247c9363e03ede381850
3
- size 951412614
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a7a1e7b09af10a4e60d0cff4a2e275f9b2b95df54b413fcae1cab54981c4c7e
3
+ size 951413126
Qwen3-0.6B-rk3588-w8a8_g512-opt-1-hybrid-ratio-0.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a66cd51090fc5ec8e8422d0463c8e5293b385d8e446024fd57d0e500fe2a498a
3
- size 950003654
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69a6ade2b986eb2962c965c6dbfed4066a5403a805f8a5ed382e0e5136d46e58
3
+ size 950004166
Qwen3-0.6B-rk3588-w8a8_g512-opt-1-hybrid-ratio-0.5.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a66cd51090fc5ec8e8422d0463c8e5293b385d8e446024fd57d0e500fe2a498a
3
- size 950003654
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69a6ade2b986eb2962c965c6dbfed4066a5403a805f8a5ed382e0e5136d46e58
3
+ size 950004166
Qwen3-0.6B-rk3588-w8a8_g512-opt-1-hybrid-ratio-1.0.rkllm CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a66cd51090fc5ec8e8422d0463c8e5293b385d8e446024fd57d0e500fe2a498a
3
- size 950003654
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69a6ade2b986eb2962c965c6dbfed4066a5403a805f8a5ed382e0e5136d46e58
3
+ size 950004166
README.md CHANGED
@@ -1,10 +1,8 @@
1
  ---
2
  base_model:
3
- - Qwen/Qwen3-0.6B-Base
4
- library_name: transformers
5
- license: apache-2.0
6
- license_link: https://huggingface.co/Qwen/Qwen3-0.6B/blob/main/LICENSE
7
- pipeline_tag: text-generation
8
  ---
9
  # Qwen3-0.6B-RK3588-1.2.1
10
 
@@ -16,11 +14,7 @@ Compatible with RKLLM version: 1.2.1
16
 
17
  # Original Model Card for base model, Qwen3-0.6B, below:
18
 
19
-
20
  # Qwen3-0.6B
21
- <a href="https://chat.qwen.ai/" target="_blank" style="margin: 2px;">
22
- <img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/>
23
- </a>
24
 
25
  ## Qwen3 Highlights
26
 
@@ -45,9 +39,6 @@ Qwen3 is the latest generation of large language models in Qwen series, offering
45
 
46
  For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our [blog](https://qwenlm.github.io/blog/qwen3/), [GitHub](https://github.com/QwenLM/Qwen3), and [Documentation](https://qwen.readthedocs.io/en/latest/).
47
 
48
- > [!TIP]
49
- > If you encounter significant endless repetitions, please refer to the [Best Practices](#best-practices) section for optimal sampling parameters, and set the ``presence_penalty`` to 1.5.
50
-
51
  ## Quickstart
52
 
53
  The code of Qwen3 has been in the latest Hugging Face `transformers` and we advise you to use the latest version of `transformers`.
@@ -105,23 +96,21 @@ print("thinking content:", thinking_content)
105
  print("content:", content)
106
  ```
107
 
108
- For deployment, you can use `sglang>=0.4.6.post1` or `vllm>=0.8.5` or to create an OpenAI-compatible API endpoint:
109
- - SGLang:
110
- ```shell
111
- python -m sglang.launch_server --model-path Qwen/Qwen3-0.6B --reasoning-parser qwen3
112
- ```
113
  - vLLM:
114
  ```shell
115
  vllm serve Qwen/Qwen3-0.6B --enable-reasoning --reasoning-parser deepseek_r1
116
  ```
117
-
118
- For local use, applications such as Ollama, LMStudio, MLX-LM, llama.cpp, and KTransformers have also supported Qwen3.
 
 
119
 
120
  ## Switching Between Thinking and Non-Thinking Mode
121
 
122
  > [!TIP]
123
- > The `enable_thinking` switch is also available in APIs created by SGLang and vLLM.
124
- > Please refer to our documentation for [SGLang](https://qwen.readthedocs.io/en/latest/deployment/sglang.html#thinking-non-thinking-modes) and [vLLM](https://qwen.readthedocs.io/en/latest/deployment/vllm.html#thinking-non-thinking-modes) users.
125
 
126
  ### `enable_thinking=True`
127
 
@@ -219,7 +208,7 @@ if __name__ == "__main__":
219
  print(f"Bot: {response_3}")
220
  ```
221
 
222
- > [!NOTE]
223
  > For API compatibility, when `enable_thinking=True`, regardless of whether the user uses `/think` or `/no_think`, the model will always output a block wrapped in `<think>...</think>`. However, the content inside this block may be empty if thinking is disabled.
224
  > When `enable_thinking=False`, the soft switches are not valid. Regardless of any `/think` or `/no_think` tags input by the user, the model will not generate think content and will not include a `<think>...</think>` block.
225
 
 
1
  ---
2
  base_model:
3
+ - Qwen/Qwen3-0.6B
4
+ tags:
5
+ - unsloth
 
 
6
  ---
7
  # Qwen3-0.6B-RK3588-1.2.1
8
 
 
14
 
15
  # Original Model Card for base model, Qwen3-0.6B, below:
16
 
 
17
  # Qwen3-0.6B
 
 
 
18
 
19
  ## Qwen3 Highlights
20
 
 
39
 
40
  For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our [blog](https://qwenlm.github.io/blog/qwen3/), [GitHub](https://github.com/QwenLM/Qwen3), and [Documentation](https://qwen.readthedocs.io/en/latest/).
41
 
 
 
 
42
  ## Quickstart
43
 
44
  The code of Qwen3 has been in the latest Hugging Face `transformers` and we advise you to use the latest version of `transformers`.
 
96
  print("content:", content)
97
  ```
98
 
99
+ For deployment, you can use `vllm>=0.8.5` or `sglang>=0.4.5.post2` to create an OpenAI-compatible API endpoint:
 
 
 
 
100
  - vLLM:
101
  ```shell
102
  vllm serve Qwen/Qwen3-0.6B --enable-reasoning --reasoning-parser deepseek_r1
103
  ```
104
+ - SGLang:
105
+ ```shell
106
+ python -m sglang.launch_server --model-path Qwen/Qwen3-0.6B --reasoning-parser deepseek-r1
107
+ ```
108
 
109
  ## Switching Between Thinking and Non-Thinking Mode
110
 
111
  > [!TIP]
112
+ > The `enable_thinking` switch is also available in APIs created by vLLM and SGLang.
113
+ > Please refer to [our documentation](https://qwen.readthedocs.io/) for more details.
114
 
115
  ### `enable_thinking=True`
116
 
 
208
  print(f"Bot: {response_3}")
209
  ```
210
 
211
+ > **Note**
212
  > For API compatibility, when `enable_thinking=True`, regardless of whether the user uses `/think` or `/no_think`, the model will always output a block wrapped in `<think>...</think>`. However, the content inside this block may be empty if thinking is disabled.
213
  > When `enable_thinking=False`, the soft switches are not valid. Regardless of any `/think` or `/no_think` tags input by the user, the model will not generate think content and will not include a `<think>...</think>` block.
214
 
added_tokens.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</think>": 151668,
3
+ "</tool_call>": 151658,
4
+ "</tool_response>": 151666,
5
+ "<think>": 151667,
6
+ "<tool_call>": 151657,
7
+ "<tool_response>": 151665,
8
+ "<|box_end|>": 151649,
9
+ "<|box_start|>": 151648,
10
+ "<|endoftext|>": 151643,
11
+ "<|file_sep|>": 151664,
12
+ "<|fim_middle|>": 151660,
13
+ "<|fim_pad|>": 151662,
14
+ "<|fim_prefix|>": 151659,
15
+ "<|fim_suffix|>": 151661,
16
+ "<|im_end|>": 151645,
17
+ "<|im_start|>": 151644,
18
+ "<|image_pad|>": 151655,
19
+ "<|object_ref_end|>": 151647,
20
+ "<|object_ref_start|>": 151646,
21
+ "<|quad_end|>": 151651,
22
+ "<|quad_start|>": 151650,
23
+ "<|repo_name|>": 151663,
24
+ "<|video_pad|>": 151656,
25
+ "<|vision_end|>": 151653,
26
+ "<|vision_pad|>": 151654,
27
+ "<|vision_start|>": 151652
28
+ }
chat_template.jinja ADDED
@@ -0,0 +1,97 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0].role == 'system' %}
4
+ {{- messages[0].content + '\n\n' }}
5
+ {%- endif %}
6
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
7
+ {%- for tool in tools %}
8
+ {{- "\n" }}
9
+ {{- tool | tojson }}
10
+ {%- endfor %}
11
+ {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
12
+ {%- else %}
13
+ {%- if messages[0].role == 'system' %}
14
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
15
+ {%- endif %}
16
+ {%- endif %}
17
+ {%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
18
+ {%- for forward_message in messages %}
19
+ {%- set index = (messages|length - 1) - loop.index0 %}
20
+ {%- set message = messages[index] %}
21
+ {%- set tool_start = '<tool_response>' %}
22
+ {%- set tool_start_length = tool_start|length %}
23
+ {%- set start_of_message = message.content[:tool_start_length] %}
24
+ {%- set tool_end = '</tool_response>' %}
25
+ {%- set tool_end_length = tool_end|length %}
26
+ {%- set start_pos = (message.content|length) - tool_end_length %}
27
+ {%- if start_pos < 0 %}
28
+ {%- set start_pos = 0 %}
29
+ {%- endif %}
30
+ {%- set end_of_message = message.content[start_pos:] %}
31
+ {%- if ns.multi_step_tool and message.role == "user" and not(start_of_message == tool_start and end_of_message == tool_end) %}
32
+ {%- set ns.multi_step_tool = false %}
33
+ {%- set ns.last_query_index = index %}
34
+ {%- endif %}
35
+ {%- endfor %}
36
+ {%- for message in messages %}
37
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
38
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
39
+ {%- elif message.role == "assistant" %}
40
+ {%- set content = message.content %}
41
+ {%- set reasoning_content = '' %}
42
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
43
+ {%- set reasoning_content = message.reasoning_content %}
44
+ {%- else %}
45
+ {%- if '</think>' in message.content %}
46
+ {%- set content = (message.content.split('</think>')|last).lstrip('\n') %}
47
+ {%- set reasoning_content = (message.content.split('</think>')|first).rstrip('\n') %}
48
+ {%- set reasoning_content = (reasoning_content.split('<think>')|last).lstrip('\n') %}
49
+ {%- endif %}
50
+ {%- endif %}
51
+ {%- if loop.index0 > ns.last_query_index %}
52
+ {%- if loop.last or (not loop.last and reasoning_content) %}
53
+ {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
54
+ {%- else %}
55
+ {{- '<|im_start|>' + message.role + '\n' + content }}
56
+ {%- endif %}
57
+ {%- else %}
58
+ {{- '<|im_start|>' + message.role + '\n' + content }}
59
+ {%- endif %}
60
+ {%- if message.tool_calls %}
61
+ {%- for tool_call in message.tool_calls %}
62
+ {%- if (loop.first and content) or (not loop.first) %}
63
+ {{- '\n' }}
64
+ {%- endif %}
65
+ {%- if tool_call.function %}
66
+ {%- set tool_call = tool_call.function %}
67
+ {%- endif %}
68
+ {{- '<tool_call>\n{"name": "' }}
69
+ {{- tool_call.name }}
70
+ {{- '", "arguments": ' }}
71
+ {%- if tool_call.arguments is string %}
72
+ {{- tool_call.arguments }}
73
+ {%- else %}
74
+ {{- tool_call.arguments | tojson }}
75
+ {%- endif %}
76
+ {{- '}\n</tool_call>' }}
77
+ {%- endfor %}
78
+ {%- endif %}
79
+ {{- '<|im_end|>\n' }}
80
+ {%- elif message.role == "tool" %}
81
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
82
+ {{- '<|im_start|>user' }}
83
+ {%- endif %}
84
+ {{- '\n<tool_response>\n' }}
85
+ {{- message.content }}
86
+ {{- '\n</tool_response>' }}
87
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
88
+ {{- '<|im_end|>\n' }}
89
+ {%- endif %}
90
+ {%- endif %}
91
+ {%- endfor %}
92
+ {%- if add_generation_prompt %}
93
+ {{- '<|im_start|>assistant\n' }}
94
+ {%- if enable_thinking is defined and enable_thinking is false %}
95
+ {{- '<think>\n\n</think>\n\n' }}
96
+ {%- endif %}
97
+ {%- endif %}
config.json CHANGED
@@ -4,7 +4,6 @@
4
  ],
5
  "attention_bias": false,
6
  "attention_dropout": 0.0,
7
- "bos_token_id": 151643,
8
  "eos_token_id": 151645,
9
  "head_dim": 128,
10
  "hidden_act": "silu",
@@ -17,14 +16,16 @@
17
  "num_attention_heads": 16,
18
  "num_hidden_layers": 28,
19
  "num_key_value_heads": 8,
 
20
  "rms_norm_eps": 1e-06,
21
  "rope_scaling": null,
22
  "rope_theta": 1000000,
23
  "sliding_window": null,
24
  "tie_word_embeddings": true,
25
  "torch_dtype": "bfloat16",
26
- "transformers_version": "4.51.0",
 
27
  "use_cache": true,
28
  "use_sliding_window": false,
29
  "vocab_size": 151936
30
- }
 
4
  ],
5
  "attention_bias": false,
6
  "attention_dropout": 0.0,
 
7
  "eos_token_id": 151645,
8
  "head_dim": 128,
9
  "hidden_act": "silu",
 
16
  "num_attention_heads": 16,
17
  "num_hidden_layers": 28,
18
  "num_key_value_heads": 8,
19
+ "pad_token_id": 151654,
20
  "rms_norm_eps": 1e-06,
21
  "rope_scaling": null,
22
  "rope_theta": 1000000,
23
  "sliding_window": null,
24
  "tie_word_embeddings": true,
25
  "torch_dtype": "bfloat16",
26
+ "transformers_version": "4.52.0.dev0",
27
+ "unsloth_fixed": true,
28
  "use_cache": true,
29
  "use_sliding_window": false,
30
  "vocab_size": 151936
31
+ }
generation_config.json CHANGED
@@ -1,13 +1,14 @@
1
  {
2
- "bos_token_id": 151643,
3
- "do_sample": true,
4
- "eos_token_id": [
5
- 151645,
6
- 151643
7
- ],
8
- "pad_token_id": 151643,
9
- "temperature": 0.6,
10
- "top_k": 20,
11
- "top_p": 0.95,
12
- "transformers_version": "4.51.0"
13
- }
 
 
1
  {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "max_length": 40960,
9
+ "pad_token_id": 151654,
10
+ "temperature": 0.6,
11
+ "top_k": 20,
12
+ "top_p": 0.95,
13
+ "transformers_version": "4.52.0.dev0"
14
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|vision_pad|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
tokenizer_config.json CHANGED
@@ -227,13 +227,15 @@
227
  "<|video_pad|>"
228
  ],
229
  "bos_token": null,
230
- "chat_template": "{%- if tools %}\n {{- '<|im_start|>system\\n' }}\n {%- if messages[0].role == 'system' %}\n {{- messages[0].content + '\\n\\n' }}\n {%- endif %}\n {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n {%- for tool in tools %}\n {{- \"\\n\" }}\n {{- tool | tojson }}\n {%- endfor %}\n {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n {%- if messages[0].role == 'system' %}\n {{- '<|im_start|>system\\n' + messages[0].content + '<|im_end|>\\n' }}\n {%- endif %}\n{%- endif %}\n{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}\n{%- for message in messages[::-1] %}\n {%- set index = (messages|length - 1) - loop.index0 %}\n {%- if ns.multi_step_tool and message.role == \"user\" and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}\n {%- set ns.multi_step_tool = false %}\n {%- set ns.last_query_index = index %}\n {%- endif %}\n{%- endfor %}\n{%- for message in messages %}\n {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) %}\n {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n {%- elif message.role == \"assistant\" %}\n {%- set content = message.content %}\n {%- set reasoning_content = '' %}\n {%- if message.reasoning_content is defined and message.reasoning_content is not none %}\n {%- set reasoning_content = message.reasoning_content %}\n {%- else %}\n {%- if '</think>' in message.content %}\n {%- set content = message.content.split('</think>')[-1].lstrip('\\n') %}\n {%- set reasoning_content = message.content.split('</think>')[0].rstrip('\\n').split('<think>')[-1].lstrip('\\n') %}\n {%- endif %}\n {%- endif %}\n {%- if loop.index0 > ns.last_query_index %}\n {%- if loop.last or (not loop.last and reasoning_content) %}\n {{- '<|im_start|>' + message.role + '\\n<think>\\n' + reasoning_content.strip('\\n') + '\\n</think>\\n\\n' + content.lstrip('\\n') }}\n {%- else %}\n {{- '<|im_start|>' + message.role + '\\n' + content }}\n {%- endif %}\n {%- else %}\n {{- '<|im_start|>' + message.role + '\\n' + content }}\n {%- endif %}\n {%- if message.tool_calls %}\n {%- for tool_call in message.tool_calls %}\n {%- if (loop.first and content) or (not loop.first) %}\n {{- '\\n' }}\n {%- endif %}\n {%- if tool_call.function %}\n {%- set tool_call = tool_call.function %}\n {%- endif %}\n {{- '<tool_call>\\n{\"name\": \"' }}\n {{- tool_call.name }}\n {{- '\", \"arguments\": ' }}\n {%- if tool_call.arguments is string %}\n {{- tool_call.arguments }}\n {%- else %}\n {{- tool_call.arguments | tojson }}\n {%- endif %}\n {{- '}\\n</tool_call>' }}\n {%- endfor %}\n {%- endif %}\n {{- '<|im_end|>\\n' }}\n {%- elif message.role == \"tool\" %}\n {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n {{- '<|im_start|>user' }}\n {%- endif %}\n {{- '\\n<tool_response>\\n' }}\n {{- message.content }}\n {{- '\\n</tool_response>' }}\n {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n {{- '<|im_end|>\\n' }}\n {%- endif %}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|im_start|>assistant\\n' }}\n {%- if enable_thinking is defined and enable_thinking is false %}\n {{- '<think>\\n\\n</think>\\n\\n' }}\n {%- endif %}\n{%- endif %}",
231
  "clean_up_tokenization_spaces": false,
232
  "eos_token": "<|im_end|>",
233
  "errors": "replace",
234
- "model_max_length": 131072,
235
- "pad_token": "<|endoftext|>",
 
 
236
  "split_special_tokens": false,
237
  "tokenizer_class": "Qwen2Tokenizer",
238
- "unk_token": null
239
- }
 
 
227
  "<|video_pad|>"
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": 40960,
235
+ "pad_token": "<|vision_pad|>",
236
+ "padding_side": "left",
237
  "split_special_tokens": false,
238
  "tokenizer_class": "Qwen2Tokenizer",
239
+ "unk_token": null,
240
+ "chat_template": "{%- if tools %}\n {{- '<|im_start|>system\\n' }}\n {%- if messages[0].role == 'system' %}\n {{- messages[0].content + '\\n\\n' }}\n {%- endif %}\n {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n {%- for tool in tools %}\n {{- \"\\n\" }}\n {{- tool | tojson }}\n {%- endfor %}\n {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n {%- if messages[0].role == 'system' %}\n {{- '<|im_start|>system\\n' + messages[0].content + '<|im_end|>\\n' }}\n {%- endif %}\n{%- endif %}\n{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}\n{%- for forward_message in messages %}\n {%- set index = (messages|length - 1) - loop.index0 %}\n {%- set message = messages[index] %}\n {%- set tool_start = '<tool_response>' %}\n {%- set tool_start_length = tool_start|length %}\n {%- set start_of_message = message.content[:tool_start_length] %}\n {%- set tool_end = '</tool_response>' %}\n {%- set tool_end_length = tool_end|length %}\n {%- set start_pos = (message.content|length) - tool_end_length %}\n {%- if start_pos < 0 %}\n {%- set start_pos = 0 %}\n {%- endif %}\n {%- set end_of_message = message.content[start_pos:] %}\n {%- if ns.multi_step_tool and message.role == \"user\" and not(start_of_message == tool_start and end_of_message == tool_end) %}\n {%- set ns.multi_step_tool = false %}\n {%- set ns.last_query_index = index %}\n {%- endif %}\n{%- endfor %}\n{%- for message in messages %}\n {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) %}\n {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n {%- elif message.role == \"assistant\" %}\n {%- set content = message.content %}\n {%- set reasoning_content = '' %}\n {%- if message.reasoning_content is defined and message.reasoning_content is not none %}\n {%- set reasoning_content = message.reasoning_content %}\n {%- else %}\n {%- if '</think>' in message.content %}\n {%- set content = (message.content.split('</think>')|last).lstrip('\\n') %}\n {%- set reasoning_content = (message.content.split('</think>')|first).rstrip('\\n') %}\n {%- set reasoning_content = (reasoning_content.split('<think>')|last).lstrip('\\n') %}\n {%- endif %}\n {%- endif %}\n {%- if loop.index0 > ns.last_query_index %}\n {%- if loop.last or (not loop.last and reasoning_content) %}\n {{- '<|im_start|>' + message.role + '\\n<think>\\n' + reasoning_content.strip('\\n') + '\\n</think>\\n\\n' + content.lstrip('\\n') }}\n {%- else %}\n {{- '<|im_start|>' + message.role + '\\n' + content }}\n {%- endif %}\n {%- else %}\n {{- '<|im_start|>' + message.role + '\\n' + content }}\n {%- endif %}\n {%- if message.tool_calls %}\n {%- for tool_call in message.tool_calls %}\n {%- if (loop.first and content) or (not loop.first) %}\n {{- '\\n' }}\n {%- endif %}\n {%- if tool_call.function %}\n {%- set tool_call = tool_call.function %}\n {%- endif %}\n {{- '<tool_call>\\n{\"name\": \"' }}\n {{- tool_call.name }}\n {{- '\", \"arguments\": ' }}\n {%- if tool_call.arguments is string %}\n {{- tool_call.arguments }}\n {%- else %}\n {{- tool_call.arguments | tojson }}\n {%- endif %}\n {{- '}\\n</tool_call>' }}\n {%- endfor %}\n {%- endif %}\n {{- '<|im_end|>\\n' }}\n {%- elif message.role == \"tool\" %}\n {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n {{- '<|im_start|>user' }}\n {%- endif %}\n {{- '\\n<tool_response>\\n' }}\n {{- message.content }}\n {{- '\\n</tool_response>' }}\n {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n {{- '<|im_end|>\\n' }}\n {%- endif %}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|im_start|>assistant\\n' }}\n {%- if enable_thinking is defined and enable_thinking is false %}\n {{- '<think>\\n\\n</think>\\n\\n' }}\n {%- endif %}\n{%- endif %}"
241
+ }