Wing Lian 9bf854e59c Phi update 202311 (#876) ...

* add phi modeling from hf

* update for packing and use new modeling class for phi

* update e2e tests for phi to use new model name

* update example phi to also use new phi model name

* use AutoModelForCausalLM for phi lora since sample packing isn't supported

2023-11-17 12:47:17 -05:00

..

cerebras

don't compile deepspeed or bitsandbytes from source (#837)

2023-11-08 19:49:55 -05:00

code-llama

don't compile deepspeed or bitsandbytes from source (#837)

2023-11-08 19:49:55 -05:00

falcon

don't compile deepspeed or bitsandbytes from source (#837)

2023-11-08 19:49:55 -05:00

gptj

don't compile deepspeed or bitsandbytes from source (#837)

2023-11-08 19:49:55 -05:00

jeopardy-bot

fix eval_steps to be a sane default (#797)

2023-10-27 22:36:30 -04:00

llama-2

various bugfixes (#856)

2023-11-15 12:23:18 -05:00

mistral

don't compile deepspeed or bitsandbytes from source (#837)

2023-11-08 19:49:55 -05:00

mpt-7b

fix eval_steps to be a sane default (#797)

2023-10-27 22:36:30 -04:00

openllama-3b

don't compile deepspeed or bitsandbytes from source (#837)

2023-11-08 19:49:55 -05:00

phi

Phi update 202311 (#876)

2023-11-17 12:47:17 -05:00

pythia

fix eval_steps to be a sane default (#797)

2023-10-27 22:36:30 -04:00

pythia-12b

simplify by removing duplicate base_model_config (#772)

2023-10-23 01:42:38 -04:00

redpajama

fix eval_steps to be a sane default (#797)

2023-10-27 22:36:30 -04:00

replit-3b

fix eval_steps to be a sane default (#797)

2023-10-27 22:36:30 -04:00

xgen-7b

don't compile deepspeed or bitsandbytes from source (#837)

2023-11-08 19:49:55 -05:00