Ömer Karışman

Results 33 comments of Ömer Karışman

@bfirsh @zeke Is there any new information about Conda support in Cog?

https://github.com/huggingface/diffusers/blob/9a23c53e4f0a1c428b84fa983f322fab6d33779c/src/diffusers/schedulers/scheduling_dpmsolver_multistep.py#L363-L371 timesteps are being overwritten here. Is this intended?

I believe that manually setting timesteps implies manual control over sigmas as well. It would be helpful if the scheduler could ignore the `use_karras_sigmas` or `use_lu_lambdas` options without generating an...

@te4336 I have teslas and i dont even get the initialization and stratum messages. Did you do something different to make it work with Tesla?

I'm very curious about that too. Based on the paper I'm assuming they introduced the control tags during fine tuning on a sample protein family (lysozymes in the case of...

@isamu-isozaki Thank you for working on this PR. @ArthurZucker instead of assuming the eos_token is the last (by id) or relying on the config, isn't it better to look the...

Below you can find example img2img sweeps with strength values from 0.5 to 1.0 comparing the current implementation in diffusers and the expected results when we use the correct timesteps:...

@fkjkey for a naive temporary solution you can change the lines: https://github.com/huggingface/diffusers/blob/279de3c3ffedcb1394518a8f1c950fa30f272390/src/diffusers/pipelines/stable_diffusion_xl/pipeline_stable_diffusion_xl_img2img.py#L1240-L1246 with: ``` timesteps, num_inference_steps = retrieve_timesteps(self.scheduler, num_inference_steps, device, timesteps, strength=strength) # timesteps, num_inference_steps = self.get_timesteps( # num_inference_steps, #...

Hi @standardAI 1. please see my response above. The diffusers code doesn’t use the strength parameter right now as not all schedulers support it 2. Diffusers make that calculation in...

Looking at the link @standardAI shared, the strength parameter is documented as > The strength and num_inference_steps parameters are related because strength determines the number of noise steps to add....