December 2024_xiangyu474_module1commit

[xiangyu474]

December 2024 Student Submission

Module Completed

• Module 1: Instruction Tuning
• Module 2: Preference Alignment
• Module 3: Parameter-efficient Fine-tuning
• Module 4: Evaluation
• Module 5: Vision-language Models
• Module 6: Synthetic Datasets
• Module 7: Inference
• Module 8: Deployment

Changes Made

Describe what you've done in this PR:

1. In 1_instruction_tuning\student_examples\xiangyu474\chat_templates_example.ipynb, I implemented 2 process_dataset functions.
2. In 1_instruction_tuning\student_examples\xiangyu474\sft_finetuning_example.ipynb, I fine tuned the base model using "everyday-conversations" dataset.

Notebooks Added/Modified

List any notebooks you've added or modified:

• Added new example in 1_instruction_tuning\student_examples\xiangyu474\chat_templates_example.ipynb
• Added new example in 1_instruction_tuning\student_examples\xiangyu474\sft_finetuning_example.ipynb
• Modified existing notebook with additional examples
• Added documentation or comments

Checklist

• I have read the module materials
• My code runs without errors
• I have pushed models and datasets to the huggingface hub
• My PR is based on the december-2024 branch

Questions or Discussion Points

Add any questions you have or points you'd like to discuss:

1. In sft_finetuning_example.ipynb, the training section was taking a long time. To address this, I modified the SFTConfig parameters, reducing max_steps from 1000 to 750. However, the resulting model output quality declined—it couldn't generate a proper haiku as requested.

Additional Notes

Any other information that might be helpful for reviewers:

[burtenshaw]

Nice work @xiangyu474 !

Would you like to take part in peer review? If so, mention me on a PR from another student, review it, and I'll get a student to review yours.

[nevernever69]

@burtenshaw this pr looks good read the files.