models

#1
by iafun - opened
Owner
β€’
edited Jul 15

As long as the original author doesn't forbid us from reproducing it on an external site, it's fine.
https://huggingface.co/John6666/vendo-pony-realistic-v12-sdxl
https://huggingface.co/John6666/lustify-sdxl-nsfwsfw-v1-sdxl

yeah thank you john for the upload !

Unfortunately i manage to find on civitai that fullreal model was deleted by his author because of new law rules in some states....and in the comments you can read people reacting to law trend and i wonder if in the future we don't have anymore that freedom to use AI model like it was at its beginning or AI model will be very restricted in some way in their use

i loved so much fullreal model, can't believe its discontinued

I see. I thought it didn't make sense, but it was a matter of state law.
It seems that the political situation is likely to become more and more regulated in the near future.
I can't do anything about it, but it is sad.

yeah thank you john for the upload !

Unfortunately i manage to find on civitai that fullreal model was deleted by his author because of new law rules in some states....and in the comments you can read people reacting to law trend and i wonder if in the future we don't have anymore that freedom to use AI model like it was at its beginning or AI model will be very restricted in some way in their use

i loved so much fullreal model, can't believe its discontinued

what was the name of the full real model? many models are being taken down but i saved some

yeah thank you john for the upload !

Unfortunately i manage to find on civitai that fullreal model was deleted by his author because of new law rules in some states....and in the comments you can read people reacting to law trend and i wonder if in the future we don't have anymore that freedom to use AI model like it was at its beginning or AI model will be very restricted in some way in their use

i loved so much fullreal model, can't believe its discontinued

what was the name of the full real model? many models are being taken down but i saved some

hello minaiosu, it was fullyREALXL_v10Perfect10n : https://civitai.com/models/227059/fullyrealxl

you can find fullrealxl v9 on hf. Unfortunately, i don't think you can upload it on Hf because the owner of this model don't want anymore its model to be used...

yeah thank you john for the upload !

Unfortunately i manage to find on civitai that fullreal model was deleted by his author because of new law rules in some states....and in the comments you can read people reacting to law trend and i wonder if in the future we don't have anymore that freedom to use AI model like it was at its beginning or AI model will be very restricted in some way in their use

i loved so much fullreal model, can't believe its discontinued

what was the name of the full real model? many models are being taken down but i saved some

hello minaiosu, it was fullyREALXL_v10Perfect10n : https://civitai.com/models/227059/fullyrealxl

you can find fullrealxl v9 on hf. Unfortunately, i don't think you can upload it on Hf because the owner of this model don't want anymore its model to be used...

can you tell me with what word should i search fullrealxlv9.0VividReal?

yeah thank you john for the upload !

Unfortunately i manage to find on civitai that fullreal model was deleted by his author because of new law rules in some states....and in the comments you can read people reacting to law trend and i wonder if in the future we don't have anymore that freedom to use AI model like it was at its beginning or AI model will be very restricted in some way in their use

i loved so much fullreal model, can't believe its discontinued

what was the name of the full real model? many models are being taken down but i saved some

hello minaiosu, it was fullyREALXL_v10Perfect10n : https://civitai.com/models/227059/fullyrealxl

you can find fullrealxl v9 on hf. Unfortunately, i don't think you can upload it on Hf because the owner of this model don't want anymore its model to be used...

fullyREALXL_v10Perfect10n its on hf too, search reica models

yes the model is there but i don't know how to upload it in order to make a space :(

you have fullyREALXL_v9VividReal but you can't upload it? i can't find it here in hf, i found all fullyrealxl from 1 to 6 in my drive but i dont have 7,8,9.. just downloaded 10 from hf

can you make a space with fullrealXL_v10 or upload it as a model with different name in order to prevent it from being removed

can you make a space with fullrealXL_v10 or upload it as a model with different name in order to prevent it from being removed

uploaded a backup model as 1010p, can you share v9 as a model instead of a space with name 0909p? 7 and 8 will be welcome too
https://huggingface.co/minaiosu/1010p/tree/main

for fullreal v9, i just duplicated this space : https://huggingface.co/GraydientPlatformAPI/fullyreal9-try2/tree/main

can you make a space with fullrealXL_v10 or upload it as a model with different name in order to prevent it from being removed

uploaded a backup model as 1010p, can you share v9 as a model instead of a space with name 0909p? 7 and 8 will be welcome too
https://huggingface.co/minaiosu/1010p/tree/main

i can't make it work as a space because i think that the model has to be upload with file organization

for fullreal v9, i just duplicated this space : https://huggingface.co/GraydientPlatformAPI/fullyreal9-try2/tree/main

cant convert it to full safetensor as it is not a bin,its already converted but not compressed, what a pain that space / model translation issues still exist

do you use a third party program in order to make work safetensor then ? which one ?

do you use a third party program in order to make work safetensor then ? which one ?

im using ComfyUI with compressed full safetensors and a 4090, i rarely use spaces

do you use a third party program in order to make work safetensor then ? which one ?

im using ComfyUI with compressed full safetensors and a 4090, i rarely use spaces

interesting is it an online tool ? do i need a good graphic card ?

do you use a third party program in order to make work safetensor then ? which one ?

im using ComfyUI with compressed full safetensors and a 4090, i rarely use spaces

interesting is it an online tool ? do i need a good graphic card ?

https://www.comfy.org/ its a full open source tool, right now you just have to download and install from the github, https://github.com/comfyanonymous/ComfyUI/releases/tag/v0.0.1, its gpu friendly

@John6666 mr john smith by any chance do you have these models or can you convert them to full working compressed safetensors?

https://huggingface.co/GraydientPlatformAPI/fullyreal9-try2/tree/main
https://huggingface.co/GraydientPlatformAPI/fullyreal9-xl

I found the means itself, but I haven't made the script yet, so you'll have to wait.
https://github.com/huggingface/diffusers/blob/main/scripts/convert_diffusers_to_original_sdxl.py

thank you very much john, im really having a hard time converting and restoring diffusers, hugging face should have the option to generate or auto-generate a safetensor file from the diffusers model_index, but sadly it doesnt, this will help a lot i will wait

i tried comfyui but i have error when i push Queue prompt button :

Error while deserializing header: HeaderTooLarge

File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\execution.py", line 151, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\execution.py", line 81, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\execution.py", line 74, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\nodes.py", line 518, in load_checkpoint
out = comfy.sd.load_checkpoint_guess_config(ckpt_path, output_vae=True, output_clip=True, embedding_directory=folder_paths.get_folder_paths("embeddings"))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\comfy\sd.py", line 482, in load_checkpoint_guess_config
sd = comfy.utils.load_torch_file(ckpt_path)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\comfy\utils.py", line 15, in load_torch_file
sd = safetensors.torch.load_file(ckpt, device=device.type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\python_embeded\Lib\site-packages\safetensors\torch.py", line 311, in load_file
with safe_open(filename, framework="pt", device=device) as f:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

i tried comfyui but i have error when i push Queue prompt button :

Error while deserializing header: HeaderTooLarge

File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\execution.py", line 151, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\execution.py", line 81, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\execution.py", line 74, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\nodes.py", line 518, in load_checkpoint
out = comfy.sd.load_checkpoint_guess_config(ckpt_path, output_vae=True, output_clip=True, embedding_directory=folder_paths.get_folder_paths("embeddings"))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\comfy\sd.py", line 482, in load_checkpoint_guess_config
sd = comfy.utils.load_torch_file(ckpt_path)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\comfy\utils.py", line 15, in load_torch_file
sd = safetensors.torch.load_file(ckpt, device=device.type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\python_embeded\Lib\site-packages\safetensors\torch.py", line 311, in load_file
with safe_open(filename, framework="pt", device=device) as f:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

can you try john6666 v9 converted version? probably v10 was not converted properly cause has a header corruption. https://huggingface.co/John6666/safetensors_converting_test/resolve/main/fullyreal9-try2.safetensors,

i tried comfyui but i have error when i push Queue prompt button :

Error while deserializing header: HeaderTooLarge

File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\execution.py", line 151, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\execution.py", line 81, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\execution.py", line 74, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\nodes.py", line 518, in load_checkpoint
out = comfy.sd.load_checkpoint_guess_config(ckpt_path, output_vae=True, output_clip=True, embedding_directory=folder_paths.get_folder_paths("embeddings"))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\comfy\sd.py", line 482, in load_checkpoint_guess_config
sd = comfy.utils.load_torch_file(ckpt_path)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\ComfyUI\comfy\utils.py", line 15, in load_torch_file
sd = safetensors.torch.load_file(ckpt, device=device.type)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users...\Downloads\ComfyUI_windows_portable\python_embeded\Lib\site-packages\safetensors\torch.py", line 311, in load_file
with safe_open(filename, framework="pt", device=device) as f:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

can you try john6666 v9 converted version? probably v10 was not converted properly cause has a header corruption. https://huggingface.co/John6666/safetensors_converting_test/resolve/main/fullyreal9-try2.safetensors,

ok i am going to test it

Hmmm...
I'll leave the current modification script for now. Let's try to find the cause.
https://huggingface.co/John6666/safetensors_converting_test/resolve/main/convert_repo_to_safetensors.py

It just occurred to me that we could do something with A1111 WebUI to repair Checkpoint.

i have a typeerror : failed to fetch, in the dos window i can see that comfyui get in pause status and ask me to push a button to go on but when i push a button, i get disconnected from comfyui

i have a typeerror : failed to fetch, in the dos window i can see that comfyui get in pause status and ask me to push a button to go on but when i push a button, i get disconnected from comfyui

probably a memory x connection issue, header and fetch problems are related with 10+ gb models

It just occurred to me that we could do something with A1111 WebUI to repair Checkpoint.

yes i think a1111 webui is the best way to test and fix this error, Error while deserializing header: HeaderTooLarge with a safetensor file, safetensor header is too big. Maximum supported size is 25000000 bytes.

i transformed it back to a primitive ckpt to test from 0 with a1111 https://huggingface.co/minaiosu/1010p/blob/main/1010pf.ckpt

It seems that 'position_ids' is lost when converted to Diffusers format, so there is no way to fix this.
I wonder if the header or metadata is the same...?
I guess unet and text encoder would work.

I got the conversion script to work with HF's Spaces.
The conversion itself takes a minute to a few minutes, but the download will take a few hours.
But it has the advantage of not needing to download locally.
https://huggingface.co/spaces/John6666/convert_repo_to_safetensors

I got the conversion script to work with HF's Spaces.
The conversion itself takes a minute to a few minutes, but the download will take a few hours.
But it has the advantage of not needing to download locally.
https://huggingface.co/spaces/John6666/convert_repo_to_safetensors

It works amazingly well. It was totally needed, thank you so much. I converted a Diffusers project to compare with the original safetensors, and they work the same. For some reason, the download is a bit slow, but it works. I just tested fullyREALXL_v10Perfect10n in a cloud service, and it's better than v9 and v8 - probably the most realistic model I have seen. Now we just need a healthy Diffusers folder, sadly most are gone.

Good. It seems to work.
I was slightly worried about the position_ids matter.

The download is slow because it is Spaces.
We could have an automatic upload mechanism to some fast models or datasets repo, but that would require me or your HF WRITE TOKEN, which would complicate things a bit.
It's not hard to make, I'll try it in my spare time. Or you can just modify it on your own. πŸ˜€

As for the specific model, we'd have to get it from somewhere anyway...

good work guys

unfortunately i cannot use comfyui, my config too low.... so if you find a solution that allow me to be able to use fullyrealv10 i will be happy to hear it.

Thank you john for uploading all the last checkpoint models !

Hello John,

Can you upload last update of this model V45 : https://civitai.com/models/49463/am-i-real ?

there is a model called twistedreality but seems to work with a different background named pixart, is it something that can work here ?

Thank you !

Owner
β€’
edited Jul 18

I am not familiar with SD 1.5 conversions, so I cannot guarantee the quality of the work.
https://huggingface.co/John6666/am-i-real-v45-sd15/

I am not familiar with SD 1.5 conversions, so I cannot guarantee the quality of the work.
https://huggingface.co/John6666/am-i-real-v45-sd15/

Thank you John !

Sign up or log in to comment