foundation-model-stack fms-extras issues

foundation-model-stack / fms-extras

Apache License 2.0

20 stars 9 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

tp world_size fix

#43 sahilsuneja1 closed 2 months ago
1
tp world_size fix

#42 sahilsuneja1 closed 2 months ago
1
Issues regarding changes incoming from the foundation-model-stack/gptq_bigcode PR branch

#41 cyang49 opened 2 months ago
2
new torch version fails mypy for process_kernel of ir.FallbackKernel

#40 JRosenkranz opened 2 months ago
0
[Draft] Incorporate vllm 0.5.5 kernels

#39 cyang49 opened 3 months ago
0
Speculator wt init fix

#38 sahilsuneja1 closed 2 months ago
2
Enable TP for paged_gpt_bigcode

#37 tdoublep closed 4 months ago
0
Add weight tying and input scaling to MLPSpeculator

#36 sahilsuneja1 closed 4 months ago
2
Add another LayerNorm to MLPSpeculator

#35 sahilsuneja1 closed 5 months ago
1
Update paged_llama.py for granite-3b-code

#34 sahilsuneja1 closed 5 months ago
0
Fused MLP in adapters

#33 JRosenkranz closed 6 months ago
2
Update paged_gpt_bigcode.py

#32 sahilsuneja1 closed 6 months ago
0
Update paged_speculative_inference.py

#31 sahilsuneja1 closed 6 months ago
0
Paged GPTBigCode Support

#30 JRosenkranz closed 5 months ago
2
Update HF dependencies

#29 ani300 closed 6 months ago
0
Move TP and adapter to new APIs for PagedLlama

#28 ani300 closed 7 months ago
0
Incorporate suggested changes from TGI PR

#27 daviswer closed 6 months ago
0
Make torch versioning slightly stricter

#26 tdoublep closed 7 months ago
0
Add speculator weight tying

#25 daviswer closed 4 months ago
0
added llama3 variants

#24 JRosenkranz closed 7 months ago
1
Updated Convention for speculator architecture/variant naming

#23 JRosenkranz closed 7 months ago
0
Fixed formatting in README

#22 JRosenkranz closed 7 months ago
0
Added information about repo in readme

#21 JRosenkranz closed 7 months ago
0
removed another fuseable weights error that was still in the repo

#20 JRosenkranz closed 7 months ago
1
removed fusable weights as they have been removed in latest version of fms

#19 JRosenkranz closed 7 months ago
0
Code LLaMA 13B Variant

#18 JRosenkranz closed 7 months ago
0
Speculative sampling

#17 daviswer opened 7 months ago
1
GPTBigCode 20b speculator variant

#16 JRosenkranz opened 7 months ago
0
up torch version compatibility

#15 JRosenkranz closed 7 months ago
0
updated fms-extras to fms>=0.0.4

#14 JRosenkranz closed 7 months ago
0
Loading/Saving Huggingface MLPSpeculator

#13 JRosenkranz closed 7 months ago
0
A "simple" v0 speculative sampling approach

#12 daviswer opened 8 months ago
0
Calico TP Embedding/Head fix

#11 JRosenkranz opened 8 months ago
0
Speculative Generation e2e

#10 JRosenkranz closed 8 months ago
0
Paged llama model

#9 JRosenkranz closed 8 months ago
0
Paged Attention KVCacheManager

#8 JRosenkranz closed 8 months ago
0
Paged Attention + Speculative Decoding support

#7 JRosenkranz closed 8 months ago
0
Add Speculator Architecture

#6 daviswer closed 9 months ago
1
Remove mypy ignore annotations

#5 afrittoli closed 10 months ago
0
configure isort

#4 nairbv closed 10 months ago
0
add github workflows

#3 nairbv closed 10 months ago
1
added initial setup.py and additions to gitignore

#2 JRosenkranz closed 10 months ago
0
initial commit of calico model with tests - including safetensors adapter support

#1 JRosenkranz closed 10 months ago
0