ultranationalism
- Joined on Oct 16, 2023
Loading Heatmap…
ultranationalism synced commits to transormer2d-variants at ultranationalism/diffusers from mirror
cb3bf76750 legacy mixin classes./
e1df77ee1e Use HF_TOKEN env var in CI (#7993)
- Compare 14 commits »
16 minutes ago
ultranationalism synced commits to main at ultranationalism/diffusers from mirror
fdb1baa05c [docs] VideoProcessor (#7965) * fix? * fix? * fix
- Compare 2 commits »
16 minutes ago
ultranationalism synced commits to feat/check-doc-listing at ultranationalism/diffusers from mirror
6529ee67ec Make VAE compatible to torch.compile() (#7984) make VAE compatible to torch.compile() Co-authored-by: YiYi Xu <yixu310@gmail.com>
df2bc5ef28 fix: Fixed few `docstrings` according to the Google Style Guide (#7717) Fixed few docstrings according to the Google Style Guide.
a7bf77fc28 Passing `cross_attention_kwargs` to `StableDiffusionInstructPix2PixPipeline` (#7961) * Update pipeline_stable_diffusion_instruct_pix2pix.py Add `cross_attention_kwargs` to `__call__` method of `StableDiffusionInstructPix2PixPipeline`, which are passed to UNet. * Update documentation for pipeline_stable_diffusion_instruct_pix2pix.py * Update docstring * Update docstring * Fix typing import
- Compare 17 commits »
16 minutes ago
ultranationalism synced commits to feat-save-sharded-ckpt at ultranationalism/diffusers from mirror
- Compare 9 commits »
16 minutes ago
ultranationalism synced commits to contact-email at ultranationalism/diffusers from mirror
16 minutes ago
ultranationalism synced commits to main at ultranationalism/ai00_rwkv_server from mirror
aafcb433f6 Fix workflow.
- Compare 6 commits »
22 hours ago
ultranationalism synced commits to main at ultranationalism/diffusers from mirror
0f0defdb65 [docs] add doc for PixArtSigmaPipeline (#7857) * 1. add doc for PixArtSigmaPipeline; --------- Co-authored-by: Sayak Paul <spsayakpaul@gmail.com> Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> Co-authored-by: Guillaume LEGENDRE <glegendre01@gmail.com> Co-authored-by: Álvaro Somoza <asomoza@users.noreply.github.com> Co-authored-by: Bagheera <59658056+bghira@users.noreply.github.com> Co-authored-by: bghira <bghira@users.github.com> Co-authored-by: Hyoungwon Cho <jhw9811@korea.ac.kr> Co-authored-by: yiyixuxu <yixu310@gmail.com> Co-authored-by: Tolga Cangöz <46008593+standardAI@users.noreply.github.com> Co-authored-by: Philip Pham <phillypham@google.com>
19df9f3ec0 Update pipeline_controlnet_inpaint_sd_xl.py (#7983)
- Compare 7 commits »
23 hours ago
ultranationalism synced commits to transormer2d-variants at ultranationalism/diffusers from mirror
6c60e430ee Consistent SDXL Controlnet callback tensor inputs (#7958) * make _callback_tensor_inputs consistent between sdxl pipelines * forgot this one * fix failing test * fix test_components_function * fix controlnet inpaint tests
- Compare 4 commits »
2 days ago
ultranationalism synced commits to main at ultranationalism/diffusers from mirror
2 days ago
ultranationalism synced commits to feat-save-sharded-ckpt at ultranationalism/diffusers from mirror
1221b28eac Fix AttributeError in train_lcm_distill_lora_sdxl_wds.py (#7923) Fix conditional teacher model check in train_lcm_distill_lora_sdxl_wds.py Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
746f603b20 Fix the text tokenizer name in logger warning of PixArt pipelines (#7912) Fix CLIP to T5 in logger warning
- Compare 8 commits »
2 days ago
ultranationalism created NPU type debugging task ultra202405081138341
4 days ago
ultranationalism pushed to master at ultranationalism/mindone
89b0b5c19b Refactor opensora pku (#495) * init * update videoae gen_video.sh * fix import error * update diffuser * do not inherit diffuser * update * update * different video shape * allow grid * allow grid * refine logging * update shell * inference on video dir * update * correct dataset name * save func error fix * updates * save data shape bs=1 * init empty files * gradio server * improve logging * avoid initiating actionvation funcs * t5 updates * add diffusion from stdit * add clip text encoder * get text encoder init.py * lpips * train causal vae * correct argparse names * example caption * correct loading path * revise precision api * changes * eval scripts * opensora ae-videobase modules * diffusion model train and infer * train shell script * adjust structure * adjust folder structure * create placeholder * align arg names * align reconstruct names * align train vae args * align diffusion * t5 revise * update t5 * update clip * allow vae input ckpt dir * update sample t2v * correct dataset * correct load path * load latte from ckpt * update sample t2v * allow index prompt embedding * allow sample image * change train common arg names * update train t2v * fix loading error * fix loading error * ckpt loading * updates * fix error * correct tokenizer mask * correct tokenizer mask * updates * requirements * revert activations * revert activations * fix typo * change precision name * cast to fp32 before asnumpy * force fp32 up/downsample in bf16 * float() to ms.float32 * change default infer precision to bf16 * correct typo * precision setting vae t5 * revise name * remove unnecessary * different t5 encoder * dropout use p not keep_prob * fix error * no need vae scale in decoding * update readme * fix save error * revise pipeline output shape * change lpips loading path * t5 force to graph mode * all inference under kbk mode * video to fp32 * allow time chunk in t2v sample * fix error * use fp16 for training now * scatter modules * fix sample typo * revise loading * allow set t5 and vae dtype * change sample loading model order * enable tiling * default enable tiling * save original rgb * allow set fps * pipeline ms.float32 * change sample method * pndm scheduler timesteps to Tensor * allow text embedding * vae bf16 examples * updates * init * update videoae gen_video.sh * fix import error * update diffuser * do not inherit diffuser * update * update shell * inference on video dir * save func error fix * updates * save data shape bs=1 * init empty files * gradio server * add diffusion from stdit * get text encoder init.py * train causal vae * correct argparse names * adjust structure * adjust folder structure * align reconstruct names * align train vae args * correct dataset * update train t2v * ckpt loading * revert activations * fix typo * change precision name * cast to fp32 before asnumpy * force fp32 up/downsample in bf16 * float() to ms.float32 * change default infer precision to bf16 * correct typo * precision setting vae t5 * revise name * remove unnecessary * different t5 encoder * dropout use p not keep_prob * fix error * no need vae scale in decoding * update readme * fix save error * revise pipeline output shape * change lpips loading path * t5 force to graph mode * all inference under kbk mode * video to fp32 * allow time chunk in t2v sample * fix error * use fp16 for training now * scatter modules * fix sample typo * revise loading * allow set t5 and vae dtype * change sample loading model order * enable tiling * default enable tiling * save original rgb * allow set fps * pipeline ms.float32 * change sample method * pndm scheduler timesteps to Tensor * allow text embedding * vae bf16 examples * updates * revert unexpected changes * get attr * update demo gif * adjust vae demo size * update demos * remove normalize extend --------- Co-authored-by: Samit <285365963@qq.com>
1576a62b63 feat(diffusers): sdxl_img2img, sdxl_inpaint, sd_img2img pipelines implementation (#474) * StableDiffusionXLImg2ImgPipeline implementation * StableDiffusionImg2ImgPipeline implementation * StableDiffusionXLInpaintPipeline implementation * Review fixes * linters fix * Line length and additional linters fixes
d008a71d67 update sdxl readme (#461) * update sdxl readme * fix CI pylint * refresh ci
1a10ba6133 add diffusers schedulers (#490)
- Compare 13 commits »
4 days ago
ultranationalism pushed to master at ultranationalism/mindone
- Compare 3 commits »
4 days ago