We introduce
SmallThinker-3B-preview
, a new model fine-tuned from the
Qwen2.5-3b-Instruct
model.
Benchmark Performance
Model
AIME24
AMC23
GAOKAO2024_I
GAOKAO2024_II
MMLU_STEM
AMPS_Hard
math_comp
Qwen2.5-3B-Instruct
6.67
45
50
35.8
59.8
-
-
SmallThinker
16.667
57.5
64.2
57.1
68.2
70
46.8
GPT-4o
9.3
-
-
-
64.2
57
50
Limitation: Due to SmallThinker's current limitations in instruction following, for math_comp we adopt a more lenient evaluation method where only correct answers are required, without constraining responses to follow the specified AAAAA format.
Intended Use Cases
SmallThinker is designed for the following use cases:
Edge Deployment:
Its small size makes it ideal for deployment on resource-constrained devices.
Draft Model for QwQ-32B-Preview:
SmallThinker can serve as a fast and efficient draft model for the larger QwQ-32B-Preview model. From my test, in llama.cpp we can get 70% speedup (from 40 tokens/s to 70 tokens/s).
Training Details
The model was trained using 8 H100 GPUs with a global batch size of 16. The specific configuration is as follows:
The SFT (Supervised Fine-Tuning) process was conducted in two phases:
First Phase:
Used only the PowerInfer/QWQ-LONGCOT-500K dataset
Trained for 1.5 epochs
Second Phase:
Combined training with PowerInfer/QWQ-LONGCOT-500K and PowerInfer/LONGCOT-Refine datasets
Continued training for an additional 2 epochs
Limitations & Disclaimer
Please be aware of the following limitations:
Language Limitation:
The model has only been trained on English-language datasets, hence its capabilities in other languages are still lacking.
Limited Knowledge:
Due to limited SFT data and the model's relatively small scale, its reasoning capabilities are constrained by its knowledge base.
Unpredictable Outputs:
The model may produce unexpected outputs due to its size and probabilistic generation paradigm. Users should exercise caution and validate the model's responses.
Repetition Issue:
The model tends to repeat itself when answering high-difficulty questions. Please increase the
repetition_penalty
to mitigate this issue.
Runs of QuantFactory SmallThinker-3B-Preview-GGUF on huggingface.co
2.1K
Total runs
0
24-hour runs
0
3-day runs
0
7-day runs
0
30-day runs
More Information About SmallThinker-3B-Preview-GGUF huggingface.co Model
SmallThinker-3B-Preview-GGUF huggingface.co
SmallThinker-3B-Preview-GGUF huggingface.co is an AI model on huggingface.co that provides SmallThinker-3B-Preview-GGUF's model effect (), which can be used instantly with this QuantFactory SmallThinker-3B-Preview-GGUF model. huggingface.co supports a free trial of the SmallThinker-3B-Preview-GGUF model, and also provides paid use of the SmallThinker-3B-Preview-GGUF. Support call SmallThinker-3B-Preview-GGUF model through api, including Node.js, Python, http.
SmallThinker-3B-Preview-GGUF huggingface.co is an online trial and call api platform, which integrates SmallThinker-3B-Preview-GGUF's modeling effects, including api services, and provides a free online trial of SmallThinker-3B-Preview-GGUF, you can try SmallThinker-3B-Preview-GGUF online for free by clicking the link below.
QuantFactory SmallThinker-3B-Preview-GGUF online free url in huggingface.co:
SmallThinker-3B-Preview-GGUF is an open source model from GitHub that offers a free installation service, and any user can find SmallThinker-3B-Preview-GGUF on GitHub to install. At the same time, huggingface.co provides the effect of SmallThinker-3B-Preview-GGUF install, users can directly use SmallThinker-3B-Preview-GGUF installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
SmallThinker-3B-Preview-GGUF install url in huggingface.co: