[Fixed] AttributeError: module 'torch' has no attribute '_utils' – Pytorch
Quick Fix: Ensure your Pytorch is updated to the latest version (0.1.2 or higher) or … Read more
[Fixed] Llama-cpp-python on AWS Sagemaker – Failed building wheel for llama-cpp-python – Python
Quick Fix: Install Visual C++ on your local machine or in the cloud environment. Ensure … Read more
Fine-tuning a pre-trained LLM for question-answering – Fine-tuning
Quick Fix: Conduct a zero-shot approach and fine-tune the T5 model on a different yet … Read more
[Fixed] ModuleNotFoundError: No module named 'transformers.models.mmbt' – How to fix it? – Python
The Problem: When user tries to import the ‘ClassificationModel’ class from the ‘simpletransformers.classification’ module, an … Read more
Splitting dataset into Train, Test and Validation using HuggingFace Datasets functions – Python
Quick Fix: You can split your dataset into train, test, and validation sets using the … Read more
[Solved] Hugging Face Transformers – trust_remote_code not working – Python
Quick Fix: Install the latest versions of transformers, accelerate, bitsandbytes, and einops using pip. Restart … Read more
[Solved] How does one use accelerate with the hugging face (HF) trainer? – Accelerate
Quick Fix: To use accelerate with the hugging face (HF) Trainer, follow these steps: Import … Read more
Loading checkpoint shards takes too long – Huggingface-transformers
Quick Fix: When using the save_pretrained() method, set safe_serialization=True to resolve the lengthy loading time … Read more
[Solved] How does one set the pad token correctly (not to eos) during fine-tuning to avoid model not predicting EOS? – Machine-learning
Quick Fix: To train on only the first occurrence of the eos token while masking … Read more
How to adapt LLaMA v2 model to less than 7B parameters? – Huggingface-transformers
Quick Fix: Reduce the model size by adjusting configuration hyperparameters, such as the ‘hidden_size’ and … Read more
[Solved] How do I slim down SBERT's sentencer-transformer library? – Python
Quick Fix: To slim down SBERT’s sentencer-transformer library when running on CPU, use the official … Read more
How to directly load fine-tuned model like Alpaca-Lora (PeftModel()) from the local files instead of load it from huggingface models? – Huggingface
Quick Fix: Use the PeftConfig object to load the configuration of the local fine-tuned Peft … Read more
Target modules for applying PEFT / LoRA on different models – Peft
Quick Fix: To apply PEFT or LoRA to different models, firstly you need to determine … Read more
[Solved] Huggingface SFT for completion only not working – Huggingface
Quick Fix: Ensure you include the formatting_func function and consider removing the dataset_text_field argument, as … Read more
[Fixed] ValueError: Tokenizer class LlamaTokenizer does not exist or is not currently imported – Llama
Quick Fix: Upgrade your transformers library by running pip install –upgrade transformers. This command will … Read more