Spaces:
Runtime error
Runtime error
Commit History
-
Update requirements.txt
Verified signaturefc1850e -
try to load gemma again
John Ho5a0f6f1 -
testing fps for internvl3 processor
John Ho3feaae8 -
added do_sample to generate
John Ho9137c51 -
make sure temperature is float
John Hoa792463 -
skipping the use of gemma model for now
John Ho96a7d4d -
added temp and testing gemma
John Ho035a7ef -
updated case matching
John Hoe13ff04 -
testing model quantization
John Ho6ec3c10 -
skip loading of internvl3 8b model
John Ho25dfae5 -
adding more internvl3 models
John Hob89bc96 -
debugging internvl3
John Hod9bd1e8 -
debugging internvl3
John Ho3eaf3ec -
debugging internvl3
John Ho4dc5aed -
fixed push logic in the last step
John Hoc5ab6e1 -
updated to not generate requirements if it already exists
John Hoca46eec -
added different inference code for internvl3
John Hod673ad7 -
added intern video model
John Ho15ef0c9 -
added control for fps and max tokens
John Hoff0b093 -
updated app to load multiple models
John Ho4361fd1
-
Update requirements.txt
Verified signatureba43302 -
Update requirements.txt
Verified signature2de01bc
-
pinning transformer to try to fix flash attention error
Verified signature1aab8b2 -
added in extra requirements
Verified signature5b5395e -
switch back to previous app
Verified signature2592317
-
attempt to slim down req
Verified signature3555eab -
Update requirements.txt
Verified signature200f657 -
Update requirements.txt
Verified signature802b23c -
chore: update requirements.txt [auto-generated by CI]
github-actions[bot]d5f0d6d -
updated torch dependency
John Hoe6a0ef4 -
trying newer torch versions
John Ho110f151 -
trying a different inference script
John Ho6dd8fb2 -
trying a different inference script
John Ho1df8e73 -
try quantization again
John Ho88958c8 -
try quantization again
John Ho4e1e198 -
trying to load model and processors outside space decorator
John Hoa83f12f -
trying model quantization
John Hoa7fd61f -
trying model quantization
John Hod9d1598 -
fixing issue with device map for the inputs
John Hoc7e712e -
added low_cpu_mem_usage and move input to device also
John Hoc697b34 -
added low_cpu_mem_usage and move input to device also
John Ho8edc124 -
testing more efficient model loading
John Hob3db9ce -
testing more efficient model loading
John Hof18bd0f -
pinning transformers version
John Hobd916de
-
make sure DTYPE is used
John Ho8b3dcea -
testing cuda for processor
John Ho1679d51 -
added back in input.to(device)
John Hoce0e222 -
make flash attention an input
John Hof10889a -
try use flash attention again
John Hof87fafd -
trying to set device outsite of spaces.GPU decorator
John Ho2a9891d