Modelpublic

ruwan/open-llama-sharded-3GB-7B-alpaca-vmware

sourceHugging Faceapache-2.0updated 3y agoView on Hugging Face

0likes20downloads

Model Card

note : use original open llama tokenizer

model_path = 'ruwan/open-llama-sharded-1GB-7B-alpaca-vmware'

tokenizer = LlamaTokenizer.frompretrained("openlm-research/openllama7b") model = LlamaForCausalLM.frompretrained( modelpath, torchdtype=torch.float16, device_map='auto' )