Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / sayakpaul/diffusers-torchao issues and pull requests

#44 - Flux Schnell not compiled

Issue - State: closed - Opened by christopher5106 9 days ago - 1 comment

#43 - sparsify for autoquant ?

Issue - State: closed - Opened by christopher5106 22 days ago - 1 comment

#42 - Combing AoT with hotswapping adapters

Issue - State: open - Opened by nom about 1 month ago - 1 comment

#41 - slow inference time when use aot to cogvideo5b

Issue - State: closed - Opened by DamonsJ 2 months ago - 10 comments

#40 - 'FakeTensor' object has no attribute '_quantized_linear_op'

Issue - State: closed - Opened by Revive624 3 months ago - 8 comments

#39 - add a guide on ahead-of-time compilation

Pull Request - State: closed - Opened by sayakpaul 3 months ago - 1 comment

#38 - Inference is fast but saving is slow.

Issue - State: closed - Opened by Leommm-byte 4 months ago - 4 comments

#37 - qkv_fuse_projections() fails with torchao quantized Flux2DTransformerModel

Issue - State: closed - Opened by ngaloppo 4 months ago - 6 comments

#36 - pipe.enable_sequential_cpu_offload()

Issue - State: closed - Opened by al-swaiti 4 months ago - 4 comments

#35 - higher batch size details.

Pull Request - State: closed - Opened by sayakpaul 5 months ago - 1 comment

#34 - gradio error- for inference

Issue - State: closed - Opened by jadechoghari 5 months ago - 13 comments

#33 - fpx_weight_only quantization + torch.compile on CogVideoX

Pull Request - State: closed - Opened by a-r-r-o-w 5 months ago - 3 comments

#32 - How should a quantized + compiled model be serialized?

Issue - State: closed - Opened by sayakpaul 5 months ago - 29 comments

#31 - cleanup utility

Pull Request - State: closed - Opened by sayakpaul 5 months ago - 2 comments

#30 - Use with lora

Issue - State: closed - Opened by akedia 5 months ago - 2 comments

#29 - Keep getting `failed to autoquant AQInt8DynamicallyQuantizedLinearWeight`

Issue - State: closed - Opened by kilimchoi 5 months ago - 3 comments

#28 - run_inference function taking too long time

Issue - State: closed - Opened by vatsal2473 5 months ago - 1 comment

#27 - Update benchmark_image.py

Pull Request - State: closed - Opened by bonlime 5 months ago

#26 - ComfyUI implementation

Issue - State: closed - Opened by pandayummy 5 months ago - 1 comment

#25 - Attribute "weight" already exists

Issue - State: closed - Opened by nachoal 5 months ago - 4 comments

#24 - Share requirements.txt file for this repo

Issue - State: closed - Opened by Tushar-ml 5 months ago - 1 comment

#23 - fpx_weight_only quantization

Pull Request - State: closed - Opened by a-r-r-o-w 5 months ago - 2 comments

#22 - Fix CogVideoX speedup numbers

Pull Request - State: closed - Opened by a-r-r-o-w 5 months ago - 3 comments

#21 - CogVideoX benchmarks with fp8dqrow

Pull Request - State: closed - Opened by a-r-r-o-w 5 months ago - 4 comments

#20 - test fp8dqrow

Pull Request - State: closed - Opened by sayakpaul 5 months ago

#19 - Issue with torchao version maybe?

Issue - State: closed - Opened by gradjitta 5 months ago - 10 comments

#18 - Add a note on sequential cpu offload for Cog

Pull Request - State: closed - Opened by a-r-r-o-w 5 months ago

#17 - Smol analysis

Pull Request - State: closed - Opened by sayakpaul 5 months ago

#16 - add flux table and smol analysis

Pull Request - State: closed - Opened by sayakpaul 5 months ago

#15 - refactor readme iii to include flux.1 dev results

Pull Request - State: closed - Opened by sayakpaul 5 months ago

#14 - simplify serialization and loading.

Pull Request - State: closed - Opened by sayakpaul 5 months ago - 3 comments

#13 - README refactor II

Pull Request - State: closed - Opened by sayakpaul 5 months ago

#12 - [FP8] Add fp8 configs

Pull Request - State: closed - Opened by drisspg 5 months ago - 2 comments

#11 - README part I

Pull Request - State: closed - Opened by sayakpaul 5 months ago

#10 - torchao-diffusers image & video model benchmarks

Pull Request - State: closed - Opened by a-r-r-o-w 5 months ago

#9 - Using quantize_to_float8 results in "Only support dtype kwarg for autocast"

Issue - State: closed - Opened by a-r-r-o-w 6 months ago - 3 comments

#8 - `torchao.autoquant` does not work with CogVideox

Issue - State: closed - Opened by a-r-r-o-w 6 months ago - 3 comments

#7 - Initial results with CogVideoX

Issue - State: closed - Opened by a-r-r-o-w 6 months ago - 3 comments

#6 - Error with FP8 compilation

Issue - State: closed - Opened by sayakpaul 6 months ago - 6 comments

#5 - Training example with FP8?

Issue - State: closed - Opened by sayakpaul 6 months ago - 14 comments

#4 - Try out FP6

Issue - State: closed - Opened by sayakpaul 6 months ago - 6 comments

#3 - End-to-end example on serializing and loading

Issue - State: closed - Opened by sayakpaul 6 months ago - 5 comments

#2 - Initial results

Issue - State: closed - Opened by sayakpaul 6 months ago - 24 comments

#1 - saprsification + dynamic int8 doesn't work

Issue - State: closed - Opened by sayakpaul 6 months ago - 14 comments