runpod / serverless-ckpt-template

🖼️ | Quickly deploy a custom RunPod Endpoint API using your own model ckpt.
MIT License
21 stars 3 forks source link

Can't add custom model #3

Open aydcery opened 11 months ago

aydcery commented 11 months ago

Hi, I am trying to add realistic vision 5.1 but not working ! I don't know what is the issue this is what I try ;

docker build --build-arg MODEL_URL=https://huggingface.co/SG161222/Realistic_Vision_V5.1_noVAE -t aydcry/rvsd:new51 . and I see this error

...
150.7 OSError: Error no file named pytorch_model.bin, tf_model.h5, model.ckpt.index or flax_model.msgpack found in directory diffusers-cache/models--SG161222--Realistic_Vision_V5.1_noVAE/snapshots/9cd4afd23ecbf0348e2c46f4ac712dbf032da73c/text_encoder.
------
Dockerfile:32
--------------------
  30 |     COPY . /opt/ckpt
  31 |
  32 | >>> RUN python3 model_fetcher.py --model_url=${MODEL_URL}
  33 |     RUN echo "Model URL: $MODEL_URL"
  34 |
--------------------
ERROR: failed to solve: process "/bin/bash -o pipefail -c python3 model_fetcher.py --model_url=${MODEL_URL}" did not complete successfully: exit code: 1
aydcery commented 11 months ago

is there a way to do it with already downloaded ckpt file?

justinmerrell commented 10 months ago

You would need to modify the code some but essentially would be skipping over the model fetch part.

kailashv7 commented 8 months ago

is there any way to make this work? I encountered the same error.

... 150.7 OSError: Error no file named pytorch_model.bin, tf_model.h5, model.ckpt.index or flax_model.msgpack found in directory diffusers-cache/models--SG161222--Realistic_Vision_V5.1_noVAE/snapshots/9cd4afd23ecbf0348e2c46f4ac712dbf032da73c/text_encoder.

Dockerfile:32

30 | COPY . /opt/ckpt 31 | 32 | >>> RUN python3 model_fetcher.py --model_url=${MODEL_URL} 33 | RUN echo "Model URL: $MODEL_URL" 34 |

ERROR: failed to solve: process "/bin/bash -o pipefail -c python3 model_fetcher.py --model_url=${MODEL_URL}" did not complete successfully: exit code: 1