jina-ai/jerboa issues and pull requests

#100 - feat: add stackoverflow dataset script

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

#100 - feat: add stackoverflow dataset script

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

#99 - save full weights and upload to hf not just adapters

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

#99 - save full weights and upload to hf not just adapters

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

#98 - feat: support dolly 15k datasets

Pull Request - State: closed - Opened by alaeddine-13 over 1 year ago

#98 - feat: support dolly 15k datasets

Pull Request - State: closed - Opened by alaeddine-13 over 1 year ago

#97 - Add automatic evaluation with gpt3

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#97 - Add automatic evaluation with gpt3

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#96 - Add dolly 15k instruction dataset

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#95 - feat: filter dataset to keep only samples ending with eos token

Pull Request - State: open - Opened by alaeddine-13 over 1 year ago

#95 - feat: filter dataset to keep only samples ending with eos token

Pull Request - State: open - Opened by alaeddine-13 over 1 year ago

#94 - fix: enforce eos token id during generation and skip special tokens when decoding

Pull Request - State: closed - Opened by alaeddine-13 over 1 year ago

#94 - fix: enforce eos token id during generation and skip special tokens when decoding

Pull Request - State: closed - Opened by alaeddine-13 over 1 year ago

#93 - Update HF models

Issue - State: closed - Opened by samsja over 1 year ago - 1 comment

#93 - Update HF models

Issue - State: closed - Opened by samsja over 1 year ago - 1 comment

#92 - fix: increase evaluation context length

Pull Request - State: open - Opened by alaeddine-13 over 1 year ago

#91 - chore: updat readme

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

#91 - chore: updat readme

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

#90 - feat: move runpod script to jina repo

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

#90 - feat: move runpod script to jina repo

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

#89 - chore: prepare for oss

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

#89 - chore: prepare for oss

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

#88 - feat: add prompter lima

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

#88 - feat: add prompter lima

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

#87 - fixing transformers version

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

#87 - fixing transformers version

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

#86 - Experiment with Lightning fabric, reproduce speed improvement from: https://lightning.ai/pages/community/finetuning-falcon-efficiently/

Issue - State: closed - Opened by sebastian-weisshaar over 1 year ago

#86 - Experiment with Lightning fabric, reproduce speed improvement from: https://lightning.ai/pages/community/finetuning-falcon-efficiently/

Issue - State: closed - Opened by sebastian-weisshaar over 1 year ago

#85 - Align Falcon 40b on alpaca-lora

Issue - State: closed - Opened by sebastian-weisshaar over 1 year ago

#84 - chore: update readme

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

#84 - chore: update readme

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

#83 - feat: select top k largest samples for `n_samples`

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

#83 - feat: select top k largest samples for `n_samples`

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

#82 - feat: add star coder

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

#82 - feat: add star coder

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

#81 - Fix memory leak

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

#81 - Fix memory leak

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

#80 - Align Falcon 7b on Lima

Issue - State: closed - Opened by sebastian-weisshaar over 1 year ago - 2 comments

#80 - Align Falcon 7b on Lima

Issue - State: closed - Opened by sebastian-weisshaar over 1 year ago - 2 comments

#79 - Align Falcon 7B on Lima

Issue - State: closed - Opened by sebastian-weisshaar over 1 year ago

#79 - Align Falcon 7B on Lima

Issue - State: closed - Opened by sebastian-weisshaar over 1 year ago

#78 - feat: add replit response

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

#78 - feat: add replit response

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

#77 - Python QA instruction tuning dataset

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

#77 - Python QA instruction tuning dataset

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

#76 - test: add save test

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

#76 - test: add save test

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

#75 - fix: model saving

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

#75 - fix: model saving

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

#74 - In our evaluation code we found a bug where the max token is just 128.

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#74 - In our evaluation code we found a bug where the max token is just 128.

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#73 - There are cases where the model is not stopping or repeats itself. We will try training for longer and see what happens

Issue - State: closed - Opened by alaeddine-13 over 1 year ago - 1 comment

#73 - There are cases where the model is not stopping or repeats itself. We will try training for longer and see what happens

Issue - State: closed - Opened by alaeddine-13 over 1 year ago - 1 comment

#72 - For Falcon, there are cases where the generation outputs an EOS token but does not stop

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#72 - For Falcon, there are cases where the generation outputs an EOS token but does not stop

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#71 - We need to train alpaca-lora on the same number of lora layers to be able to compare it to falcon 7B and understand the effect of changing from llama to falcon

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#71 - We need to train alpaca-lora on the same number of lora layers to be able to compare it to falcon 7B and understand the effect of changing from llama to falcon

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#70 - Compare code aligned model to current SOTA

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

#70 - Compare code aligned model to current SOTA

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

#69 - chore: bump transformers and all

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

#68 - Create an 8bits model for inference out of falcon 40b

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

#67 - feat: script to execute training run and stop runpod afterwards

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

#67 - feat: script to execute training run and stop runpod afterwards

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

#66 - fix: readme cli after switching to typer

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

#66 - fix: readme cli after switching to typer

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

#65 - Chore update dependency

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

#65 - Chore update dependency

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

#64 - Add lima dataset to the training pipeline

Issue - State: closed - Opened by alaeddine-13 over 1 year ago - 1 comment

#64 - Add lima dataset to the training pipeline

Issue - State: closed - Opened by alaeddine-13 over 1 year ago - 1 comment

#63 - feat: update peft ot latest main

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

#63 - feat: update peft ot latest main

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

#62 - chore: random name with model and data path name

Pull Request - State: closed - Opened by alaeddine-13 over 1 year ago - 1 comment

#62 - chore: random name with model and data path name

Pull Request - State: closed - Opened by alaeddine-13 over 1 year ago - 1 comment

#61 - feat: add type hint to load model

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

#61 - feat: add type hint to load model

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

#60 - docs: add runpod setup

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

#60 - docs: add runpod setup

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

#59 - Fix bug in save_pretrained

Issue - State: closed - Opened by samsja over 1 year ago
Labels: bug

#59 - Fix bug in save_pretrained

Issue - State: closed - Opened by samsja over 1 year ago
Labels: bug

#58 - feat: allow customizing dataset and dataset processing and support redpajamas instruct datasets

Pull Request - State: closed - Opened by alaeddine-13 over 1 year ago

#58 - feat: allow customizing dataset and dataset processing and support redpajamas instruct datasets

Pull Request - State: closed - Opened by alaeddine-13 over 1 year ago

#57 - refactor: remove useless file

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

#57 - refactor: remove useless file

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

#56 - feat: add typer and logging of config

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

#56 - feat: add typer and logging of config

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

#55 - feat: docker image

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

#54 - feat: docker container for runpod

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

#54 - feat: docker container for runpod

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

#53 - Add dockerfile to jerboa to run on runpod

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

#53 - Add dockerfile to jerboa to run on runpod

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

#52 - logs dataset in wandb

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

#52 - logs dataset in wandb

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

#51 - Pipeline training dataset refactoring

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#51 - Pipeline training dataset refactoring

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#50 - refactor: allow to work with any causal LLM including falcon

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

#50 - refactor: allow to work with any causal LLM including falcon

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

#49 - Add red pajamas instruct dataset to our pipeline

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#49 - Add red pajamas instruct dataset to our pipeline

Issue - State: closed - Opened by alaeddine-13 over 1 year ago

#48 - add redpajama 7b to our pipeline

Issue - State: closed - Opened by samsja over 1 year ago - 1 comment

#48 - add redpajama 7b to our pipeline

Issue - State: closed - Opened by samsja over 1 year ago - 1 comment

GitHub / jina-ai/jerboa issues and pull requests