Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / nebuly-ai/nebuly issues and pull requests

#352 - Forward Forward Algorithm Questions

Issue - State: open - Opened by and-rewsmith about 1 year ago - 2 comments

#351 - [DOCS] Fix broken links

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

#350 - How to generate and perform inference for an ONNX model

Issue - State: open - Opened by dneemuth about 1 year ago - 2 comments

#348 - Yolov8-Pose Model

Issue - State: open - Opened by saim212 over 1 year ago

#347 - torch2.0 support on speedster

Issue - State: open - Opened by lucasjinreal over 1 year ago - 2 comments

#345 - Use new Python API of OpenVINO

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

#343 - Evaluating accuracy of only the reward model

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

#341 - [Speedster] Optimization failed with PytorchBackendCompiler

Issue - State: open - Opened by nemeziz69 over 1 year ago - 4 comments

#340 - Standford SHP reward dataset

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

#339 - Fix wrong imports

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

#337 - [Chatllama] what's supposed to be in the Actor checkpoint dir?

Issue - State: open - Opened by StrangeTcy over 1 year ago - 3 comments

#336 - pegasus speedster notebook added

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

#335 - [chatllama]How models enable inference

Issue - State: open - Opened by yangzhipeng1108 over 1 year ago - 1 comment

#334 - Speedster refactor, add support for TPUs and AWS Inferentia chips

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

#333 - Speedster refactor, add support for TPUs and AWS Inferentia chips

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

#332 - ViT pytorch speedster notebook added

Pull Request - State: closed - Opened by VinishUchiha over 1 year ago - 7 comments

#331 - Issues with accelerate and deepspeed training

Issue - State: open - Opened by swang99 over 1 year ago - 4 comments

#330 - [ChatLLaMA] Reward Dataset Score Calculation

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

#329 - [ChatLLaMA] RL Trainer - is_deepspeed_init variable initialization

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

#328 - [Nebullvm] Add models and utils to Nebullvm SDK

Pull Request - State: closed - Opened by Telemaco019 over 1 year ago

#327 - [ChatLLaMA] DocFix - RL training command in README.md missing filename

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

#326 - Fix deepspeed multi gpu model saving

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

#325 - Support for torch 2.0

Issue - State: open - Opened by lminer over 1 year ago - 1 comment
Labels: speedster

#324 - module not found:chatllama.rlhf.dataset

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

#323 - [Chatllama] facebook/opt-350m is missing in rlhf/model_list.py

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

#322 - [chatllama]Do I need to split the llama model manully?

Issue - State: open - Opened by balcklive over 1 year ago - 2 comments

#321 - [Chatllama] Merge the datasets to create more insightful training data

Issue - State: open - Opened by PierpaoloSorbellini over 1 year ago - 2 comments
Labels: good first issue, chatllama

#320 - [Chatllama] Support Inference for trained models.

Issue - State: open - Opened by PierpaoloSorbellini over 1 year ago - 1 comment
Labels: good first issue, chatllama

#319 - [Chatllama] Evaluation Function and Loop with metrics

Issue - State: open - Opened by PierpaoloSorbellini over 1 year ago
Labels: good first issue, chatllama

#317 - [ChatLLaMA] Will there be online Demo about ChatLLaMA?

Issue - State: open - Opened by xihuai18 over 1 year ago - 1 comment

#316 - Implement deepspeed multi gpu for RL Resolves #307

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

#312 - [ChatLLaMA] RLHF Training: dimension mismatch

Issue - State: open - Opened by BigRoddy over 1 year ago - 3 comments

#311 - Add Generator and FloatTensor to torch class

Pull Request - State: closed - Opened by Telemaco019 over 1 year ago

#309 - [ChatLLaMA] Training process halted

Issue - State: open - Opened by MuffinC over 1 year ago - 3 comments

#308 - [Speedster] Would nebullvm support YoloV8 segmentation soon or at all ?

Issue - State: open - Opened by flyzxm5177 over 1 year ago - 1 comment
Labels: speedster

#307 - [ChatLLaMA] train RL on multi-gpu

Issue - State: closed - Opened by EthanChen1234 over 1 year ago - 3 comments

#306 - Refactor of models and trainers with base class for common methods

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

#305 - Update config.yaml

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

#304 - Add peft_config.yaml description

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

#303 - [ChatLLaMA] Add support for LoRA and PeFT efficient training

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

#302 - [ChatLLaMA] No GPU Detected issue

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

#301 - [Speedster] Running into "CUDA out of memory" on an A100

Issue - State: open - Opened by michaelbogdan over 1 year ago - 9 comments
Labels: speedster

#300 - Add tokenizer path to load_tokenizer for llama models

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

#299 - [ChatLLaMA] RLHF Training: Prompt too long

Issue - State: open - Opened by swang99 over 1 year ago - 8 comments
Labels: bug, chatllama

#298 - [Chatllama] KL Divergence equation

Issue - State: open - Opened by mountinyy over 1 year ago - 3 comments

#297 - Tokenizer Issue

Issue - State: closed - Opened by yradwan147 over 1 year ago - 3 comments

#295 - Add Support for PEFT fine-tuning

Issue - State: open - Opened by PierpaoloSorbellini over 1 year ago - 1 comment
Labels: chatllama

#294 - [Speedster] Fix bug when diffusers is not installed

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

#293 - [Speedster] type object ‘DummyClass’ has no attribute ‘models’

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

#291 - [Chatllama] fix typo of discounted_rewards in PPO loss

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

#290 - [Chatllama] discounted_rewards with PPO

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

#289 - RL trainig RuntimeError: CUDA error: device-side assert triggered

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

#288 - for rlhf_accelerate branch, can't run with multiGPU

Issue - State: open - Opened by balcklive over 1 year ago - 2 comments

#287 - [Nebullvm] Improve typing

Pull Request - State: closed - Opened by Telemaco019 over 1 year ago

#286 - How to inference?

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

#285 - [Speedster] Speedster usage

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

#284 - [ChatLlama] Error in the start of OPT1.3B actor pre-training

Issue - State: closed - Opened by swang99 over 1 year ago - 1 comment

#283 - Fix issue with visualization of some jupyter notebooks on github

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

#282 - [Speedster] - Example notebooks not working

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

#281 - [Chatllama] looks like REWARD MODEL donesn't support OPT

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

#273 - [Speedster] `optimize_model` function transfers the original model from the GPU to the CPU

Issue - State: open - Opened by mfumanelli over 1 year ago - 6 comments
Labels: good first issue, speedster

#272 - [Speedster] speedster usage

Issue - State: open - Opened by vinnitu over 1 year ago - 5 comments
Labels: speedster

#268 - [Chatllama] Is the training correct?

Issue - State: closed - Opened by menandro over 1 year ago - 3 comments

#266 - [chatllama]training batch_size=2 would crash

Issue - State: open - Opened by BeyonderXX over 1 year ago - 4 comments

#265 - [Chatllama] Generate Data Error on tutorial

Issue - State: closed - Opened by yicheng-2019 over 1 year ago - 2 comments

#260 - [Chatllama] Should I add a <end_of_text> at the end of sentence?

Issue - State: open - Opened by bino282 over 1 year ago - 3 comments

#254 - [Chatllama]: MultiGPU support for training

Issue - State: open - Opened by TejaGollapudi over 1 year ago - 8 comments

#251 - [Chatllama] Actor training for GPT2 while using deepspeed

Issue - State: closed - Opened by Yottaxx over 1 year ago - 8 comments

#248 - [Speedster] Fix #239

Pull Request - State: closed - Opened by Telemaco019 over 1 year ago - 1 comment

#243 - [Chatllama] Input/completion in reward training vs. RL training

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

#240 - [Chatllama] RLHF training for Actor

Issue - State: open - Opened by Vincent131499 over 1 year ago - 4 comments

#236 - [Chatllama] Actor for llama

Issue - State: closed - Opened by Vincent131499 over 1 year ago - 4 comments

#234 - [Chatllama] BUG?Assertion srcIndex < srcSelectDimSize failed.

Issue - State: open - Opened by iMountTai over 1 year ago - 14 comments

#233 - [Chatllama] support for accelerate from hf

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

#232 - [Chatllama] cannot import name 'TrainingStats' from 'utils'

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

#229 - [Chatllama] error when load dataset when use deepspeed

Issue - State: open - Opened by bino282 over 1 year ago - 5 comments

#226 - Add support for pre-trained reward models

Issue - State: open - Opened by diegofiori over 1 year ago - 2 comments
Labels: chatllama

#224 - [Chatllama] Use upvotes in Stanford dataset as a measure for reward

Issue - State: open - Opened by diegofiori over 1 year ago - 8 comments
Labels: good first issue, chatllama

#221 - [Chatllama] Add multiple sources for generating synthetic data

Issue - State: open - Opened by diegofiori over 1 year ago - 1 comment
Labels: good first issue, chatllama

#218 - [Chatllama] Add support to other open-source models

Issue - State: open - Opened by diegofiori over 1 year ago
Labels: good first issue, chatllama

#167 - Add Faster Transformer compiler for Bert

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