hf-llm-api / networks

Commit History

  1. :boom: [Fix] Remove whole token which ends with stop_sequences

    2d7c092
  2. :boom: [Fix] Remove stop_sequences from content

    8a167b0
  3. :boom: [Fix] slow tokenizer issue, and finish by stop_sequences

    8697de8

  1. :zap: [Enhance] Remove pycrypto requirements, and prettify log

    24cceb8
  2. :zap: [Enhance] Replace Crypto with hashlib

    0616d17
  3. :gem: [Feature] Enable gpt-3.5-turbo by adding proof_of_work step

    3061fdd
  4. :gem: [Feature] New ProofWorker: Used in openai auth

    c59b13c
  5. :gem: [Feature] New HfApiException: INVALID_API_KEY_ERROR

    a1d960f

  1. :zap: [Enhance] HuggingchatStreamer: logger auth messages

    862427b
  2. :zap: [Enhance] Use chrome impoersonate for cffi_requests, and upgrade to v1.2.5

    0618e58
  3. :zap: [Enhance] Use nous-mixtral-8x7b as default model

    214fb7b
  4. :zap: [Enhance] Quieter openai auth, use cffi to request hf-chat id, and console token count for exceeds

    0f710a2
  5. :boom: [Fix] Value of (delta_)content or content_type not assigned

    b40b5fc
  6. :gem: [Feature] Moduralize TokenChecker, and fix gated model repos with alternatives

    8df3985
  7. :gem: [Feature] HuggingchatStreamer: Support no-stream mode

    caedafb
  8. :gem: [Feature] HuggingchatStreamer: Enable chat_return_generator

    62d5db7
  9. :gem: [Feature] New TokenChecker: count tokens and check token limit

    64105c5
  10. :gem: [Feature] Split to Requester and Streamer, and mock chat history with messages

    c43287d
  11. :gem: [Feature] Enable chat response with get_message_id

    e4d11b8
  12. :hammer: [WIP] Enabling get_message_id

    a5ac953
  13. :gem: [Feature] HuggingchatStreamer: Build pipeline of chat_response

    55b0c51
  14. :gem: [Feature] HuggingchatStreamer: New log request and response

    c706328
  15. :gem: [Feature] HuggingchatStreamer: New get_hf_chat_id, and improve get_conversation_id

    391cdfe
  16. :gem: [Feature] New HuggingchatStreamer: get_conversation_id

    f1218fc

  1. :gem: [Feature] Enable non-stream return, and prettify logger

    397529a
  2. :boom: [Fix] Set default model

    5dc8ebf
  3. :recycle: [Refactor] Rename gpt-3.5 to gpt-3.5-turbo

    d36d623
  4. :gem: [Feature] Enable gpt-3.5 in chat_api

    cd6b52a
  5. :recycle: [Refactor] Rename message_streamer to huggingface_streamer

    3125c87
  6. :recycle: [Refactor] Use tclogger, and configure envs better

    babcd78

  1. :recycle: [Refactor] Move STOP_SEQUENCES_MAP and TOKEN_LIMIT_MAP to constants

    bf8c5bd
  2. :recycle: [Refactor] Move MODELS_MAP to constants

    3f608c6
  3. :gem: [Feature] Support new model: openchat-3.5-0106

    9f5d69c

  1. :gem: [Feature] New model enabled: gemma-7b

    77b5a47

  1. :zap: [Enhance] Set min and max for top_p

    64645f0
  2. :gem: [Feature] Add top_p and set 0.95 as default

    403b8cf
  3. :gem: [Feature] Add use_cache parameter, and set default temperature to 0.5

    e2b245b

  1. :gem: [Feature] New model enabled: nous-mixtral-8-7b

    4a28ca5

  1. :zap: [Enhance] Remove openchat-3.5 from available models

    06c8d58

  1. :boom: [Fix] inputs_tokens + max_new_tokens must <= 32768

    fa3bfb0
  2. :boom: [Fix] temperature must be strictly positive

    eb00725

  1. :boom: [Fix] Error 500 when temperature is null

    8ba223c
  2. :zap: [Enhance] Auto calculate max_tokens if not set

    1b9f698

  1. :zap: [Enhance] Remove leading whitespaces of first response

    c95d47e
  2. :pencil: [Doc] Available models: More comments on dropped models

    85354a9

  1. :gem: [Feature] MessageStreamer: Support stop sequence and auto removal

    4ba2ca6
  2. :gem: [Feature] Support call hf api with api_key via HTTP Bearer

    2da6968

  1. :gem: [Feature] Support no-stream mode with dict response

    d2b20f2

  1. :pencil: [Doc] Readme for features, and comments for references

    e820e51
  2. :zap: [Enhance] Cleaner logger output

    b96cef7