optimum-intel

optimum-intel copied to clipboard

Add JPQD evaluation notebook

3

Add JPQD evaluation notebook. Since JPQD QA takes about 12 hours to train, it doesn't make sense to do it in a notebook (if the browser crashes or the computer...

helena-intel

Try to run nueral_compressor/language_modeling, as follows. it just same as on read.me. I have 24G GPU, but cause GPU OOM. This model is only 125M, is it normal? How much...

lcw99

OVModelForCausalLM OOM

1

``` #from transformers import AutoTokenizer, BloomModel from optimum.intel.openvino import OVModelForCausalLM from transformers import AutoTokenizer, BloomModel import torch from tqdm import tqdm from time import time from time import sleep model_str...

Zjq9409

fix backward compatibility for GenerationMode import

2

# What does this PR do? fix import GenerationMode for transformers

eaidova

update codegen config for support codegen2

1

# What does this PR do? update codegen config for support codegen2, added support qwen2moe and dbrx models ## Before submitting - [ ] This PR fixes a typo or...

eaidova

Fix llama and gemma modeling patching for openvino export

1

Fix compatibility for transformers v4.41.0 cc @helena-intel @eaidova

echarlaix

Update quantized_generation_demo.ipynb

4

Fix version of torch to 2.0.1. See https://github.com/pytorch/pytorch/issues/125109 # What does this PR do? Running this notebook with current pytorch version fails. Test Platform: Windows 11 - Intel Core Ultra...

dnoliver

Update default 4bit configs

3

# What does this PR do? Contains configuration updates based on the experiments from 135227 and following PR's: - https://github.com/openvinotoolkit/openvino.genai/pull/377 - https://github.com/openvinotoolkit/openvino.genai/pull/419 ## Before submitting - [ ] This PR...

KodiaqQ

[OpenVINO] Set Left Padding For Text Generation Task

1

# What does this PR do? Set the left padding side during tokenizer conversion for text generation tasks. Fixes # (issue) ## Before submitting - [ ] This PR fixes...

apaniukov

Convert i64 input tensors to i32 for GPU plugin

11

# What does this PR do? OpenVINO GPU plugin does not support int64 natively so i64 inputs are always converted to i32. To avoid runtime conversion, updated IO tensor precision...

yeonbok