Malaysian-Qwen3-0.6B-test

Uploaded model

• Developed by: Faris-Faiz
• License: apache-2.0
• Finetuned from model : unsloth/qwen3-0.6b

Performance is trash cause the loss during training is so high, need to increase max steps:

                       Model   Accuracy   shot by_letter        category
0  Malaysian-Qwen3-0.6B-test  31.764224  0shot      True            STEM
1  Malaysian-Qwen3-0.6B-test  35.225827  0shot      True        Language
2  Malaysian-Qwen3-0.6B-test  31.743278  0shot      True  Social science
3  Malaysian-Qwen3-0.6B-test  33.197409  0shot      True          Others
4  Malaysian-Qwen3-0.6B-test  35.585893  0shot      True      Humanities
{'Social science': np.int64(6918), 'Language': np.int64(6288), 'Humanities': np.int64(4395), 'Others': np.int64(4169), 'STEM': np.int64(2443)}
Model : Malaysian-Qwen3-0.6B-test
Metric : first
Shot : 0shot
average accuracy 33.59765415272788
accuracy for STEM 31.764224314367578
accuracy for Language 35.225826972010175
accuracy for Social science 31.743278404163057
accuracy for Others 33.1974094507076
accuracy for Humanities 35.585893060295795

This qwen3 model was trained 2x faster with Unsloth and Huggingface's TRL library.

<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>