Modeling code for Mistral to use with
Nanotron
Also contains converted pretrained weights for Mistral-7B-0.1:
https://huggingface.co/mistralai/Mistral-7B-v0.1
1# Generate a config file
2python config_tiny_mistral.py
3
4# Run training
5export CUDA_DEVICE_MAX_CONNECTIONS=1 # important for some distributed operations
6torchrun --nproc_per_node=8 run_train.py --config-file config_tiny_mistral.yaml
1export CUDA_DEVICE_MAX_CONNECTIONS=1
2torchrun --nproc_per_node=1 run_generate.py --ckpt-path ./pretrained/Mistral-7B-v0.1