text-generation-inference
Adding Llava-Next (Llava 1.6) with full support.
#1709
Merged
Merged
Llava next dump.
b68fc4de
Update by abstracting away text model.
b8be0d1a
More work on the CLIP Side.
5f4b3954
Tmp dump (running on images hardcoded size.)
df4c7008
Working for TP, Llama + Mistral
6c350f2f
Adding docs.
7852a85b
drbh
commented
on 2024-04-05
Fixing integration tests ? (Failures locally).
ccbfc05d
Upgrade tests (still missing load tests for some reason).
99771cfa
Fixed load test. Bad sanitation on the router meant CUDA OOM.
39620ce2
More GPUs for more VRAM.
274b68ad
Tmp dump (sending real image for real memory offset to be computed.
215030ac
Created all the logic server side (with image download on the fly too).
2283562b
Update mt0 (not more truncating).
61821f41
drbh
commented
on 2024-04-09
Fixing select_best_resolution.
8c114e5f
Move import up.
4217ddb8
drbh
approved these changes
on 2024-04-09
Login to write a write a comment.
Login via GitHub