Joe Rocca PRO
rocca
AI & ML interests
None yet
Recent Activity
upvoted
a
paper
about 1 hour ago
XQuant: Breaking the Memory Wall for LLM Inference with KV Cache
Rematerialization
liked
a Space
2 days ago
MCG-NJU/PixNerd
liked
a model
5 days ago
ibm-granite/granite-embedding-english-r2
Organizations
chroma-unlocked-v4x-hyper-turbo-flash-r64-bf16
👍
2
1
#5 opened 22 days ago
by
mweldon
Which model to be loaded?
4
#3 opened 22 days ago
by
Skymid
FP4 quantization, INT4 crash on RTX 50XX
🔥
1
7
#2 opened 3 months ago
by
Askd234
Any chance of creating a v30 version to hold over until real support?
👍
3
9
#1 opened 3 months ago
by
Todokete
Does FlashMLA support kv cache fp8 dtype and how to enable FlashMLA ?
10
#6 opened 5 months ago
by
CharlesLincoln
Diffusers Roadmap?
👍
1
4
#5 opened 5 months ago
by
Impulse2000
70B model?
5
#1 opened about 1 year ago
by
rocca
vllm can not inter this model (other 70b gptq model are ok)
15
#1 opened over 1 year ago
by
tutu329
Is `--speculate` needed for Medusa in TGI?
2
#1 opened over 1 year ago
by
rocca
Double-hyphen in "gptq-3bit--1g-actorder_True" branch name is invalid according to TGI
1
#1 opened over 1 year ago
by
rocca
Conversion to ONNX
4
#1 opened almost 3 years ago
by
roseman
Delete the outdated rwkv-4-pile-430m-webgl.onnx
1
#4 opened almost 3 years ago
by
AXKuhta
Upload rwkv-4-pile-430m-webgl.onnx
1
#3 opened almost 3 years ago
by
AXKuhta
Upload rwkv-4-pile-430m-webgl.onnx
1
#2 opened almost 3 years ago
by
AXKuhta
Upload rwkv-4-pile-169m-webgl.onnx
1
#1 opened almost 3 years ago
by
AXKuhta
This is really exciting
2
#1 opened about 3 years ago
by
victor