Repository Issues
lm-sys/FastChat
An open platform for training, serving, and evaluating large language models. Release repo for Vicuna and Chatbot Arena.
Issues
Offen
TensorRT-LLM backend support?
enhancementgood first issue
3 Kommentare7 Reaktionen0 zugewiesene Personen
Offen
Please support fastllm
good first issue
2 Kommentare0 Reaktionen0 zugewiesene Personen
Offen
Support microsoft phi-1.5 model
good first issue
4 Kommentare0 Reaktionen0 zugewiesene Personen
Offen
[Feature request] Support loading GGUF and GGML model format
good first issue
5 Kommentare7 Reaktionen0 zugewiesene Personen
Offen
[Feature Request] Support InternLM Deploy
good first issue
2 Kommentare0 Reaktionen0 zugewiesene Personen
Offen
Need to support Baichuan2
good first issue
11 Kommentare12 Reaktionen0 zugewiesene Personen
Offen
Issues with VLLM Integration Speedup
good first issue
4 Kommentare0 Reaktionen0 zugewiesene Personen
Offen
XVERSE-13B need support!
good first issue
1 Kommentar0 Reaktionen0 zugewiesene Personen
Offen
[Feature] Do you have plan to support multimodal mode?
good first issue
1 Kommentar1 Reaktion0 zugewiesene Personen
Offen
[Feature Request] Add a ctranslate2 model worker
enhancementgood first issue
2 Kommentare1 Reaktion1 zugewiesene Person
Offen
Add model C1.2 from Character.ai?
good first issue
3 Kommentare0 Reaktionen0 zugewiesene Personen
Offen
RWKV-raven and RWKV-world have different prompt template
good first issue
3 Kommentare0 Reaktionen0 zugewiesene Personen
Offen
4 Kommentare1 Reaktion0 zugewiesene Personen
Offen
presence_penalty and repetition_penalty in completions endpoint
good first issue
1 Kommentar0 Reaktionen0 zugewiesene Personen
Offen
support for 4bit quantization from transfomer library.
enhancementgood first issue
7 Kommentare2 Reaktionen0 zugewiesene Personen
Offen
tiiuae/falcon-7b does not work on Apple M1 GPU (MPS)
good first issue
2 Kommentare1 Reaktion0 zugewiesene Personen
Offen
Support fastchat-t5-3b-v1.0 on M2 GPU model
good first issue
1 Kommentar0 Reaktionen0 zugewiesene Personen
Offen
Any plans to add AutoGPTQ as a gptq load option?
good first issue
1 Kommentar0 Reaktionen0 zugewiesene Personen
Offen
Chat Arena version of fastchat-t5-3b-v1.0 provides more refined answers than standard Huggingface model.
documentationgood first issuequestion
26 Kommentare3 Reaktionen1 zugewiesene Person
Offen
[BUG] RWKV Models are not configured to use Cuda GPU lists.
buggood first issue
2 Kommentare0 Reaktionen0 zugewiesene Personen