microsoft/TransformerCompression issues and pull requests

#184 - Support Phi3.5 and upgrade Transformers

Pull Request - State: open - Opened by LianaMikael 6 days ago

#183 - transformers dependency version upgrade

Issue - State: open - Opened by xiaoyu-work 19 days ago

#100 - Define Q matrices as trainable parameters

Pull Request - State: closed - Opened by LianaMikael 11 months ago - 1 comment

#100 - Define Q matrices as trainable parameters

Pull Request - State: closed - Opened by LianaMikael 11 months ago - 1 comment

#99 - Make Qs trainable parameters

Pull Request - State: closed - Opened by LianaMikael 11 months ago

#99 - Make Qs trainable parameters

Pull Request - State: closed - Opened by LianaMikael 11 months ago

#98 - Shift labels during finetuning and provide correct padding mask

Pull Request - State: closed - Opened by LianaMikael 11 months ago

#98 - Shift labels during finetuning and provide correct padding mask

Pull Request - State: closed - Opened by LianaMikael 11 months ago

#97 - Add tests that run experiments

Pull Request - State: closed - Opened by myshkov 11 months ago

#97 - Add tests that run experiments

Pull Request - State: closed - Opened by myshkov 11 months ago

#96 - Fix sliced model loading for HF models

Pull Request - State: closed - Opened by pashminacameron 11 months ago - 1 comment

#95 - Make perplexity evaluation general

Pull Request - State: closed - Opened by pashminacameron 11 months ago

#95 - Make perplexity evaluation general

Pull Request - State: closed - Opened by pashminacameron 11 months ago

#94 - Update wikitext2 dataloader tokenizer call

Pull Request - State: closed - Opened by pashminacameron 11 months ago - 1 comment

#94 - Update wikitext2 dataloader tokenizer call

Pull Request - State: closed - Opened by pashminacameron 11 months ago - 1 comment

#93 - Remove duplicate script parameter

Pull Request - State: closed - Opened by myshkov 11 months ago

#93 - Remove duplicate script parameter

Pull Request - State: closed - Opened by myshkov 11 months ago

#92 - Log lm_eval metrics concisely

Pull Request - State: closed - Opened by pashminacameron 11 months ago

#92 - Log lm_eval metrics concisely

Pull Request - State: closed - Opened by pashminacameron 11 months ago

#91 - Support local models

Pull Request - State: closed - Opened by pashminacameron 12 months ago

#91 - Support local models

Pull Request - State: closed - Opened by pashminacameron 12 months ago

#90 - Fix dataset handling in run_benchmark

Pull Request - State: closed - Opened by lopusz 12 months ago

#89 - Different logic for test dataloader

Pull Request - State: closed - Opened by nailimixaM 12 months ago - 1 comment

#89 - Different logic for test dataloader

Pull Request - State: closed - Opened by nailimixaM 12 months ago - 1 comment

#88 - James/unrotate

Pull Request - State: closed - Opened by jameshensman 12 months ago

#88 - James/unrotate

Pull Request - State: closed - Opened by jameshensman 12 months ago

#87 - Update lora_target_map format

Pull Request - State: closed - Opened by pashminacameron 12 months ago

#87 - Update lora_target_map format

Pull Request - State: closed - Opened by pashminacameron 12 months ago

#86 - Update README.md with hf link

Pull Request - State: closed - Opened by jameshensman 12 months ago

#86 - Update README.md with hf link

Pull Request - State: closed - Opened by jameshensman 12 months ago

#85 - Fix Padding Tokens Initialization for Llama and Phi

Pull Request - State: closed - Opened by LianaMikael 12 months ago

#85 - Fix Padding Tokens Initialization for Llama and Phi

Pull Request - State: closed - Opened by LianaMikael 12 months ago

#84 - Update README.md

Pull Request - State: closed - Opened by eltociear 12 months ago

#83 - Pass round interval param to model load

Pull Request - State: closed - Opened by myshkov 12 months ago

#83 - Pass round interval param to model load

Pull Request - State: closed - Opened by myshkov 12 months ago

#82 - Fix slicing for parallel blocks case

Pull Request - State: closed - Opened by myshkov 12 months ago

#82 - Fix slicing for parallel blocks case

Pull Request - State: closed - Opened by myshkov 12 months ago

#81 - Mistral Support

Issue - State: open - Opened by fakerybakery 12 months ago - 5 comments

#81 - Mistral Support

Issue - State: open - Opened by fakerybakery 12 months ago - 5 comments

#80 - Update README

Pull Request - State: closed - Opened by nailimixaM 12 months ago

#80 - Update README

Pull Request - State: closed - Opened by nailimixaM 12 months ago

#79 - Add more abstract properties to model adapter

Pull Request - State: closed - Opened by pashminacameron 12 months ago

#79 - Add more abstract properties to model adapter

Pull Request - State: closed - Opened by pashminacameron 12 months ago

#78 - Add config abstract property in model adapter.

Pull Request - State: closed - Opened by pashminacameron 12 months ago - 2 comments

#78 - Add config abstract property in model adapter.

Pull Request - State: closed - Opened by pashminacameron 12 months ago - 2 comments

#77 - Rename Compressible -> Compressed

Pull Request - State: closed - Opened by pashminacameron 12 months ago

#77 - Rename Compressible -> Compressed

Pull Request - State: closed - Opened by pashminacameron 12 months ago

#76 - Scrub keys from commit history

Issue - State: closed - Opened by nailimixaM 12 months ago

#75 - Add varying per layer sparsity

Pull Request - State: closed - Opened by myshkov 12 months ago

#75 - Add varying per layer sparsity

Pull Request - State: closed - Opened by myshkov 12 months ago

#74 - Run syne-tune on AMD GPUs

Pull Request - State: closed - Opened by PocherninaElena 12 months ago

#74 - Run syne-tune on AMD GPUs

Pull Request - State: closed - Opened by PocherninaElena 12 months ago

#73 - loading models is painful and not HF compatible

Issue - State: open - Opened by jameshensman almost 1 year ago

#73 - loading models is painful and not HF compatible

Issue - State: open - Opened by jameshensman almost 1 year ago

#72 - parameter count is innacurate

Issue - State: closed - Opened by jameshensman almost 1 year ago - 1 comment

#72 - parameter count is innacurate

Issue - State: closed - Opened by jameshensman almost 1 year ago - 1 comment

#71 - Cleaner model adapter interface

Pull Request - State: closed - Opened by pashminacameron about 1 year ago

#71 - Cleaner model adapter interface

Pull Request - State: closed - Opened by pashminacameron about 1 year ago

#70 - Update llama_adapter.py

Pull Request - State: closed - Opened by pashminacameron about 1 year ago

#70 - Update llama_adapter.py

Pull Request - State: closed - Opened by pashminacameron about 1 year ago

#69 - Fix layers split across GPUs when using --distribute

Pull Request - State: closed - Opened by nailimixaM about 1 year ago

#69 - Fix layers split across GPUs when using --distribute

Pull Request - State: closed - Opened by nailimixaM about 1 year ago

#68 - Update project authors

Pull Request - State: closed - Opened by jameshensman about 1 year ago

#68 - Update project authors

Pull Request - State: closed - Opened by jameshensman about 1 year ago

#67 - Expand phi-2 finetuning to include lm_head

Pull Request - State: closed - Opened by pashminacameron about 1 year ago

#67 - Expand phi-2 finetuning to include lm_head

Pull Request - State: closed - Opened by pashminacameron about 1 year ago

#66 - --distribute fails for calculating perplexity

Issue - State: closed - Opened by nailimixaM about 1 year ago
Labels: bug

#66 - --distribute fails for calculating perplexity

Issue - State: closed - Opened by nailimixaM about 1 year ago
Labels: bug

#65 - Fix SliceGPT for Llama

Pull Request - State: closed - Opened by nailimixaM about 1 year ago

#65 - Fix SliceGPT for Llama

Pull Request - State: closed - Opened by nailimixaM about 1 year ago

#64 - Update README with links to arxiv paper (when uploaded)

Issue - State: closed - Opened by nailimixaM about 1 year ago

#64 - Update README with links to arxiv paper (when uploaded)

Issue - State: closed - Opened by nailimixaM about 1 year ago

#63 - Finetuning phi-2, Bayes opt and minor bugfixes

Pull Request - State: closed - Opened by nailimixaM about 1 year ago

#62 - Initial public api

Pull Request - State: closed - Opened by msdmkats about 1 year ago - 2 comments

#62 - Initial public api

Pull Request - State: closed - Opened by msdmkats about 1 year ago - 2 comments

#61 - Use hf-token from env var or cmd arg

Pull Request - State: closed - Opened by pashminacameron about 1 year ago

#61 - Use hf-token from env var or cmd arg

Pull Request - State: closed - Opened by pashminacameron about 1 year ago

#60 - Replace <pad> with <eos> for Llama2 and Phi2

Issue - State: closed - Opened by nailimixaM about 1 year ago

#60 - Replace <pad> with <eos> for Llama2 and Phi2

Issue - State: closed - Opened by nailimixaM about 1 year ago

#59 - Pashmina/phi2 (support latest phi-2 and transformers)

Pull Request - State: closed - Opened by pashminacameron about 1 year ago

#59 - Pashmina/phi2 (support latest phi-2 and transformers)

Pull Request - State: closed - Opened by pashminacameron about 1 year ago

#58 - hf token should be retrivable from env

Issue - State: closed - Opened by jameshensman about 1 year ago - 1 comment

#57 - Pashmina/readme

Pull Request - State: closed - Opened by pashminacameron about 1 year ago

#56 - c4 dataset download fails

Issue - State: open - Opened by nailimixaM about 1 year ago - 1 comment
Labels: bug

#55 - Skip c4 test as quick fix to PR long build issue

Pull Request - State: closed - Opened by nailimixaM about 1 year ago

#54 - Fix lm eval harness for zeroshot experiments

Pull Request - State: closed - Opened by nailimixaM about 1 year ago

#53 - James/docstrings for adapter

Pull Request - State: closed - Opened by jameshensman about 1 year ago - 1 comment

#52 - Common command-line args

Issue - State: open - Opened by nailimixaM about 1 year ago
Labels: enhancement

#51 - James/bo finetune

Pull Request - State: closed - Opened by jameshensman about 1 year ago

#50 - Pashmina/phi2

Pull Request - State: closed - Opened by jameshensman about 1 year ago

#49 - llamadecoder in transformers 4.36

Issue - State: closed - Opened by jameshensman about 1 year ago - 1 comment

#48 - Slicing with batches of varying sizes and seqlens fails

Issue - State: closed - Opened by nailimixaM about 1 year ago
Labels: bug

#47 - adding and end-to-end test for opt slicing

Pull Request - State: open - Opened by jameshensman about 1 year ago - 2 comments

#46 - adding rounding option to script

Pull Request - State: closed - Opened by jameshensman about 1 year ago

#45 - Fix benchmark experiment

Pull Request - State: closed - Opened by nailimixaM about 1 year ago

#44 - Set dtype and device for all tensors in PCA test

Pull Request - State: closed - Opened by myshkov about 1 year ago

#43 - Max/finetuning

Pull Request - State: closed - Opened by nailimixaM about 1 year ago

#42 - A general interface for new models

Pull Request - State: closed - Opened by msdmkats about 1 year ago - 1 comment

#41 - Max's dataset refactor

Pull Request - State: closed - Opened by jameshensman about 1 year ago - 1 comment

#40 - adds the options to run models in fp32

Pull Request - State: closed - Opened by jameshensman about 1 year ago

GitHub / microsoft/TransformerCompression issues and pull requests