scaleapi/llm-engine

Issues

• Guided choice not respected

  #612 opened 6 months ago by Andrea-de-Varda
  2

• RetNet adaptation

  #266 opened 6 months ago by yunfeng-scale
  1

• Further reduction of pod cold start time

  #269 opened 6 months ago by yunfeng-scale
  1

• PEFT adapters with continuous batching

  #268 opened 6 months ago by yunfeng-scale
  1

• Speculative decoding

  #267 opened 6 months ago by yunfeng-scale
  1

• Investigate CUDA graphs

  #265 opened 6 months ago by yunfeng-scale
  1

• GQA for Llama 2 7B and 13B models

  #264 opened 6 months ago by yunfeng-scale
  1

• Add support for Mistral in the FineTune API

  #308 opened 6 months ago by ian-scale
  1

• Return log probabilities of the prompt

  #616 opened 6 months ago by ChiaSap
  2

• ⚠️ Notice: sunsetting the public demo of LLM Engine ⚠️

  #619 opened 6 months ago by yixu34
  0

• Guided choice with mixtral model is not honored

  #528 opened 7 months ago by KE7
  1

• Test out spot instances

  #270 opened 10 months ago by yunfeng-scale
  1

• how can we use glue dataset to llama2-7B？

  #499 opened a year ago by ZHANGJINKUI
  1

• Add Completions support for Llama 3

  #497 opened a year ago by yixu34
  1

• Fine-Tuning LLM using Local GPU and Infra

  #477 opened a year ago by Sree-abcprocure
  1

• Integrate TensorRT-LLM

  #363 opened a year ago by yunfeng-scale
  0

• Cannot pass in PEFT configs when creating a finetuning job

  #325 opened a year ago by tu-trinh
  4

• Add support for Mistral-7B in the Completions API

  #301 opened 2 years ago by yixu34
  2

• self host on runnpod

  #285 opened 2 years ago by Stealthwriter
  1

• Error: Internal Server Error: <class 'AttributeError'>: 'CreateFineTuneResponse' object has no attribute 'artifact_id'

  #272 opened 2 years ago by ashutoshrana171
  4

• Control frequency - completion

  #277 opened 2 years ago by Stealthwriter
  3

• ⚠️ LLM Engine fine-tuning maintenance ⚠️

  #233 opened 2 years ago by yixu34
  4

• Investigate Multi-Query Attention

  #171 opened 2 years ago by yixu34
  2

• [Feature Request] support InternLM

  #206 opened 2 years ago by JimmyMa99
  1

• Fine-tuning API should return immediately with a clear error message if the input is invalid

  #213 opened 2 years ago by yixu34
  0

• Surface pandas ParserError to user

  #184 opened 2 years ago by squeakymouse
  0

• Parsing Error Raised when Attempting to Run Example Code

  #146 opened 2 years ago by Arahat-Chikkatur
  5

• Bug: API calls don't work on Windows due to os.path.join

  #209 opened 2 years ago by f-linus
  1

• [Feature Request] Add additional options for text generation

  #187 opened 2 years ago by jenkspt
  4

• [Lora] Allow more Lora hyperparams

  #163 opened 2 years ago by sam-scale
  1

• [Datasets] s3 presigned url fails?

  #168 opened 2 years ago by sam-scale
  1

• Model ids =/= Fine Tune Id?

  #167 opened 2 years ago by sam-scale
  1

• Add github sidebar

  #155 opened 2 years ago by yixu34
  1

• Remove `status` and `traceback` from completion response on the server

  #159 opened 2 years ago by yixu34
  1

• Llama-2-70B support

  #166 opened 2 years ago by yixu34
  7

• Can we fine tune with data stored in local csv file?

  #221 opened 2 years ago by jaslatendresse
  9

• [Tracking] Allow wandb tracking

  #165 opened 2 years ago by sam-scale
  1

• Import completion error

  #182 opened 2 years ago by 4n9le-bot
  2

• max token length for finetune and completion endpoints on Lllama-2?

  #208 opened 2 years ago by urimerhav
  2

• FineTune.create - NotFoundError (API endpoint seems to through 404)

  #170 opened 2 years ago by f-linus
  7

• [Feature Request] Add `on_inference_ready` callback to llm-engine deployments via `Model.create`

  #193 opened 2 years ago by jenkspt
  1

• [Feature Request] Add token log probabilities to response

  #188 opened 2 years ago by jenkspt
  1

• [dev] Deploy docs through CI

  #189 opened 2 years ago by phil-scale
  0

• FineTune.get_events errors when there are no FineTune events yet

  #169 opened 2 years ago by rkaplan
  3

• Allow users to set API key without using env variables

  #156 opened 2 years ago by yixu34
  0

• Example ScienceQA fine-tuning notebook has no documentation for how to install dependencies

  #172 opened 2 years ago by rkaplan
  1

• [Tokenizer] Allow for custom tokens/tokenizer

  #164 opened 2 years ago by sam-scale
  0

• [Checkpointing] Allow users to save/use multiple checkpoints

  #162 opened 2 years ago by sam-scale
  0

• model name parameter after finetuning

  #143 opened 2 years ago by eshrager
  7

• GKE Helm deployment

  #157 opened 2 years ago by brosand
  3