huu4ontocord/mdel issues and pull requests

#69 - update cbtm inference

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

#68 - Add clustering

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

#67 - Create lora.py

Pull Request - State: open - Opened by huu4ontocord about 1 year ago

#66 - Do a small test run

Issue - State: open - Opened by mrcabbage972 over 1 year ago

#65 - Set up the training configuration

Issue - State: open - Opened by mrcabbage972 over 1 year ago

#64 - Tokenize the StarCoder dataset

Issue - State: open - Opened by mrcabbage972 over 1 year ago - 1 comment

#63 - Get all relevant data for StarCoder into LUMI

Issue - State: open - Opened by mrcabbage972 over 1 year ago

#62 - Conda environment

Pull Request - State: closed - Opened by henk717 over 1 year ago - 1 comment

#61 - Draft: Create train_cbtm_classifier.py

Pull Request - State: closed - Opened by kenhktsui over 1 year ago - 2 comments

#60 - Train 2nd batch of expert models

Issue - State: open - Opened by mrcabbage972 over 1 year ago

#59 - Add files via upload

Pull Request - State: closed - Opened by Digitous over 1 year ago - 1 comment

#58 - Create template for HF dataset config

Issue - State: open - Opened by mrcabbage972 over 1 year ago
Labels: Trainer

#57 - Investigate Expert Models Having High Perplexity

Issue - State: open - Opened by mrcabbage972 over 1 year ago - 1 comment
Labels: Evaluation

#56 - Get rid of HFArgParse

Pull Request - State: closed - Opened by jordiclive over 1 year ago - 3 comments

#55 - Perplexity Script and Notebook

Pull Request - State: closed - Opened by Stillerman over 1 year ago - 2 comments

#54 - Fixing pre-commit errors

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago

#53 - Evaluate a merged expert model's perplexity

Issue - State: open - Opened by mrcabbage972 over 1 year ago - 3 comments
Labels: Evaluation

#52 - Merge N experts notebook

Pull Request - State: closed - Opened by Stillerman over 1 year ago - 3 comments

#51 - Stabilize Training on Redmond Box

Issue - State: open - Opened by mrcabbage972 over 1 year ago
Labels: Trainer

#50 - c-btm inference

Pull Request - State: closed - Opened by NourFahmy over 1 year ago - 6 comments

#49 - Update trainer_chat.py data handling and add shell script

Pull Request - State: closed - Opened by P1ayer-1 over 1 year ago - 3 comments

#48 - Fixed report_to index bug in trainer.py

Pull Request - State: closed - Opened by P1ayer-1 over 1 year ago - 3 comments

#47 - Automatic Training Scripts for All Expert Models

Issue - State: open - Opened by kenhktsui over 1 year ago
Labels: Trainer

#46 - add DS option.

Pull Request - State: closed - Opened by jordiclive over 1 year ago

#45 - Setup separate environments on Redmond.ai box

Issue - State: open - Opened by mrcabbage972 over 1 year ago - 2 comments
Labels: Trainer

#44 - inputs_ids cast to fp16 in deeperspeed bug

Issue - State: open - Opened by mayank31398 over 1 year ago

#43 - Create eval_merges.py

Pull Request - State: closed - Opened by thedarkzeno over 1 year ago - 1 comment

#42 - Train baseline models for evaluation

Issue - State: open - Opened by huu4ontocord over 1 year ago - 10 comments
Labels: Training Runs

#41 - Training instruction followers as composable layers and expert layers

Issue - State: open - Opened by huu4ontocord over 1 year ago - 1 comment
Labels: Training Runs

#40 - Expert merging: c-BTM

Issue - State: open - Opened by mrcabbage972 over 1 year ago - 3 comments
Labels: enhancement, Merging

#39 - Update README.md to include the full list of pile subsets

Pull Request - State: closed - Opened by P1ayer-1 over 1 year ago

#38 - Update trainer.py to include wandb run in hf card

Pull Request - State: closed - Opened by P1ayer-1 over 1 year ago - 2 comments

#37 - Integrate with LLM evaluation frameworks

Issue - State: open - Opened by kenhktsui over 1 year ago - 3 comments
Labels: Evaluation

#36 - Add script for merging expert models via weight averaging

Issue - State: open - Opened by mrcabbage972 over 1 year ago - 7 comments
Labels: Merging

#35 - add multiple validation loaders and wandb init

Pull Request - State: closed - Opened by jordiclive over 1 year ago

#34 - Fix HF Hub Upload Error

Issue - State: closed - Opened by mrcabbage972 over 1 year ago

#33 - Report val loss aggregated by data origin

Issue - State: closed - Opened by mrcabbage972 over 1 year ago

#32 - Streamline training

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago

#31 - Fixed spelling mistake in train.sh

Pull Request - State: closed - Opened by P1ayer-1 over 1 year ago

#30 - Fixed spelling mistake in train.sh

Pull Request - State: closed - Opened by P1ayer-1 over 1 year ago

#29 - Using steps eval strategy

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago

#28 - Refactor train script

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago

#27 - Adding train shell script

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago

#26 - Add shell script for single layer training

Pull Request - State: closed - Opened by P1ayer-1 over 1 year ago

#25 - Dataset generation open issues

Issue - State: open - Opened by mrcabbage972 over 1 year ago

#24 - Shard pile

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago

#23 - Using the original Pile data to retain the splits

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago

#22 - Add code for uploading data to huggingface

Pull Request - State: closed - Opened by P1ayer-1 over 1 year ago

#21 - Add code for uploading data to huggingface

Pull Request - State: closed - Opened by P1ayer-1 over 1 year ago

#20 - Add code for uploading data to huggingface

Pull Request - State: closed - Opened by P1ayer-1 over 1 year ago

#19 - Update README.md

Pull Request - State: open - Opened by mrcabbage972 over 1 year ago

#18 - Fixed python naming bug in create_uspto_pile_mix.sh

Pull Request - State: closed - Opened by P1ayer-1 over 1 year ago

#17 - Process all domain data files

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago

#16 - Change the dataset mixing script to process files with a wild card

Issue - State: closed - Opened by mrcabbage972 over 1 year ago

#15 - Adding code to mix Pile and expert data

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago - 2 comments

#14 - Add code for mixing Pile and Expert data

Issue - State: closed - Opened by mrcabbage972 over 1 year ago

#13 - Adding Adalberto's HF trainer

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago

#12 - Add Single Layer HF Trainer

Issue - State: closed - Opened by mrcabbage972 over 1 year ago

#11 - Create minimal example of training on SUMMIT

Issue - State: open - Opened by mrcabbage972 over 1 year ago

#10 - Create minimal example of training on LUMI

Issue - State: open - Opened by mrcabbage972 over 1 year ago - 1 comment

#9 - 7 add pylint config

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago - 1 comment

#8 - Create pylint.yml

Pull Request - State: open - Opened by mrcabbage972 over 1 year ago

#7 - Add pylint

Issue - State: closed - Opened by mrcabbage972 over 1 year ago

#6 - Add genre classifer code to repo

Issue - State: open - Opened by mrcabbage972 over 1 year ago

#5 - Adding folder structure

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago

#4 - Establish folder structure

Issue - State: closed - Opened by mrcabbage972 over 1 year ago

#3 - Adding Resource document

Pull Request - State: closed - Opened by mrseeker over 1 year ago - 1 comment

#2 - Adding pre-commit and makefile

Pull Request - State: closed - Opened by mrcabbage972 over 1 year ago - 1 comment

#1 - Add pre-commit hooks

Issue - State: closed - Opened by mrcabbage972 over 1 year ago

GitHub / huu4ontocord/mdel issues and pull requests