simple-llm-finetuner
simple-llm-finetuner copied to clipboard
lxe
Simple UI for LLM Model Finetuning
**In trainer.py, ignore the last token is not suitable for all situations.** ```python def tokenize_sample(self, item, max_seq_length, add_eos_token=True): assert self.tokenizer is not None result = self.tokenizer( item["text"], truncation=True, max_length=max_seq_length, padding="max_length",...
This pull request introduces the `--use_cpu` option to the main script, allowing users to run training and inference on the CPU. The `--use_cpu` option can be specified when running the...
Running inference gives the following warning: ``` Loading tokenizer... The tokenizer class you load from this checkpoint is not the same type as the class this function is called from....
Hi there! I would like to know how we can run this solution in multi gpu environment for bigger models. Thank you
While I run the train in colab, this error is shown - ``` Something went wrong Connection errored out. ``` How can I solve this?
Training on T4: torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 44.00 MiB (GPU 0; 14.56 GiB total capacity; 13.25 GiB already allocated; 10.44 MiB free; 13.83 GiB reserved in...
You did an amazing job. But most people wants a cli version to quickly get started and play around. Here it is. Cheers.
I have added progressive output to the inference tab by converting the `generate` function in app.py to a Python generator and producing tokens one at at time until either the...
