LMFlow
LMFlow copied to clipboard
OptimalScale
An Extensible Toolkit for Finetuning and Inference of Large Foundation Models. Large Models for All.
Could you tell me how to use this conversation_template in the chatbot? I used a training dataset that follows the Llama-3 conversation_template, but there doesn’t seem to be an argument...
After lora finetune, I get my insurance model, can I use run_benchmark.sh script to evaluate my insurance model on my customized dataset?
Can your fine-tuning code be used for pre-training of the model? That is, the unsupervised fine-tuning stage of knowledge injection. If so, can you tell me the corresponding code file?...
I just ran Qwen2.5 with LMFlow and using the conversation template in [Link](https://optimalscale.github.io/LMFlow/examples/supported_conversation_template.html#qwen-2). The LLM can run but in the beginning of its answer, there is some weird word e.g....
bash data/download.sh all Connecting to lmflow.org (lmflow.org)|107.23.182.175|:5000... failed: Connection timed out.
How can I evaluate the MMLU dataset if LFFlow only supports text2text and textonly dataset formats, and the MMLU dataset from LISA does not fit either of these?
Fix #823 , we will merge to main later after some tests
Metadata
Owner
Metadata
An Extensible Toolkit for Finetuning and Inference of Large Foundation Models. Large Models for All.