Edit model card

INT8 T5 large finetuned on CNN DailyMail

Post-training dynamic quantization

This is an INT8 PyTorch model quantized with huggingface/optimum-intel through the usage of Intel® Neural Compressor.

The original fp32 model comes from the fine-tuned model sysresearch101/t5-large-finetuned-xsum-cnn.

Evaluation result

INT8 FP32
Accuracy (eval-rougeLsum) 29.6346 29.7451
Model size 879M 3021M

Load with optimum:

from optimum.intel import INCModelForSeq2SeqLM

model_id = "Intel/t5-large-finetuned-xsum-cnn-int8-dynamic"
int8_model = INCModelForSeq2SeqLM.from_pretrained(model_id)
Downloads last month
4
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Dataset used to train Intel/t5-large-finetuned-xsum-cnn-int8-dynamic-inc

Collection including Intel/t5-large-finetuned-xsum-cnn-int8-dynamic-inc