metadata
base_model: skt/A.X-4.0
quantized_by: tmfi-us
license: other
license_name: qwen
license_link: https://huggingface.co/skt/A.X-4.0/blob/main/LICENSE
language:
- en
- ko
pipeline_tag: text-generation
About
FP8 w8a8 dynamic quants of skt/A.X-4.0.
Used following Python script with llmcompressor to generate:
from transformers import AutoTokenizer, AutoModelForCausalLM
from llmcompressor.transformers import oneshot
from llmcompressor.modifiers.quantization import QuantizationModifier
MODEL_ID = 'skt/A.X-4.0'
model = AutoModelForCausalLM.from_pretrained(
MODEL_ID, device_map="auto", torch_dtype="auto",
)
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
# Configure the simple PTQ quantization
recipe = QuantizationModifier(
targets="Linear", scheme="FP8_DYNAMIC", ignore=["lm_head"])
# Apply the quantization algorithm.
oneshot(model=model, recipe=recipe)
# Save the model.
SAVE_DIR = MODEL_ID.split("/")[1] + "-FP8-Dynamic"
model.save_pretrained(SAVE_DIR)
tokenizer.save_pretrained(SAVE_DIR)
Quantization recipe can be found in recipe.yaml