fearnworks fearnworks
fearnworks
/ run_qlora.sh
Last active
June 2, 2023 17:50
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| #!/bin/bash | |
| # | |
| # Container source: https://github.com/OpenAccess-AI-Collective/axolotl/blob/main/docker/Dockerfile-runpod | |
| # | |
| # | |
| # To run this in RunPod with `winglian/axolotl-runpod:main-cu118-2.0.0`, set | |
| # Expose HTTP Ports (Max 10): 7860,8888 | |
| # docker command: `bash -c "curl -H 'Cache-Control: no-cache' https://raw.githubusercontent.com/utensil/llm-playground/main/scripts/entry/prepare_ax.sh -sSf | bash"` | |
| # JUPYTER_PASSWORD change to your secret | |
| # HUGGINGFACE_TOKEN change to your token from https://huggingface.co/settings/tokens |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| base_model: openlm-research/open_llama_3b_600bt_preview | |
| base_model_config: openlm-research/open_llama_3b_600bt_preview | |
| model_type: LlamaForCausalLM | |
| tokenizer_type: LlamaTokenizer | |
| load_in_8bit: false | |
| load_in_4bit: true | |
| strict: false | |
| push_dataset_to_hub: | |
| datasets: | |
| - path: teknium/GPT4-LLM-Cleaned |
fearnworks
/ config.yml
Created
May 28, 2023 18:42
This config creates a Error invalid device ordinal at line 359 in file /mnt/d/training_area/bitsandbytes/csrc/pythonInterface.c error on a A6000 single gpu run with axolotl
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| # Changing this would change the base model used for training | |
| base_model: huggyllama/llama-7b | |
| # Changing this would change the configuration used for the base model | |
| base_model_config: huggyllama/llama-7b | |
| # Changing this would change the type of model used for training | |
| model_type: LlamaForCausalLM | |
| # Changing this would change the type of tokenizer used for tokenizing text data | |
| tokenizer_type: LlamaTokenizer | |
| # Changing this to false would prevent the model from being loaded in 8-bit precision | |
| load_in_8bit: false |
NewerOlder