mobiuslabsgmbh

/

Llama-3.1-8b-instruct_4bitgs64_hqq

like 2

Follow

Mobius Labs GmbH 93

torch 2.4
transformers 4.4

import torch
from transformers import AutoTokenizer
from hqq.models.hf.base import AutoHQQHFModel
from hqq.utils.patching import *
from hqq.core.quantize import *
from hqq.utils.generation_hf import HFGenerator

#Load the model
###################################################
model_id = 'mobiuslabsgmbh/Llama-3.1-8b-instruct_4bitgs64_hqq' #no calib version
#model_id = 'mobiuslabsgmbh/Llama-3.1-8b-instruct_4bitgs64_hqq_calib' #calibrated version

compute_dtype = torch.bfloat16 #bfloat16 for torchao, float16 for bitblas
cache_dir = '.'
model = AutoHQQHFModel.from_quantized(model_id, cache_dir=cache_dir, compute_dtype=compute_dtype)
tokenizer = AutoTokenizer.from_pretrained(model_id, cache_dir=cache_dir)

quant_config = BaseQuantizeConfig(nbits=4, group_size=64, quant_scale=False, quant_zero=False, axis=1)
patch_linearlayers(model, patch_add_quant_config, quant_config)

#Use optimized inference kernels
###################################################
HQQLinear.set_backend(HQQBackend.PYTORCH)
#prepare_for_inference(model) #default backend
prepare_for_inference(model, backend="torchao_int4")
#prepare_for_inference(model, backend="bitblas") #takes a while to init...

#Generate
###################################################
#For longer context, make sure to allocate enough cache via the cache_size= parameter
gen = HFGenerator(model, tokenizer, max_new_tokens=1000, do_sample=True, compile="partial").warmup() #Warm-up takes a while

gen.generate("Write an essay about large language models", print_tokens=True)
gen.generate("Tell me a funny joke!", print_tokens=True)
gen.generate("How to make a yummy chocolate cake?", print_tokens=True)

in <cell line: 0>()
13 compute_dtype = torch.bfloat16 #bfloat16 for torchao, float16 for bitblas
14 cache_dir = '.'
---> 15 model = AutoHQQHFModel.from_quantized(model_id, cache_dir=cache_dir, compute_dtype=compute_dtype)
16 tokenizer = AutoTokenizer.from_pretrained(model_id, cache_dir=cache_dir)
17

5 frames
/usr/local/lib/python3.11/dist-packages/transformers/models/llama/configuration_llama.py in _rope_scaling_validation(self)
179
180 if not isinstance(self.rope_scaling, dict) or len(self.rope_scaling) != 2:
--> 181 raise ValueError(
182 "rope_scaling must be a dictionary with two fields, type and factor, " f"got {self.rope_scaling}"
183 )

ValueError: rope_scaling must be a dictionary with two fields, type and factor, got {'factor': 8.0, 'high_freq_factor': 4.0, 'low_freq_factor': 1.0, 'original_max_position_embeddings': 8192, 'rope_type': 'llama3'}