deepseek-ai

/

DeepSeek-R1

like 5.24k

Follow

DeepSeek 20.9k

When I use
path = 'deepseek-ai/DeepSeek-R1'
tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(path, trust_remote_code=True),

there is a bug called:

ValueError: Unknown quantization type, got fp8 - supported types are: ['awq', 'bitsandbytes_4bit', 'bitsandbytes_8bit', 'gptq', 'aqlm', 'quanto', 'eetq', 'hqq', 'compressed-tensors', 'fbgemm_fp8', 'torchao', 'bitnet']

Does anyone know how to solve it?