dalle-mini

dalle-mini copied to clipboard

Shape mismatch for commit

3

I'm trying to run the inference notebook, but I get this error while loading dall-e mini, not sure how it gets checkpoint shape to 259, I guess it should be...

Ahmedniz1

Model (~5GB) has to be redownloaded when the container is restarted

5

Each time I restart the container I have to paste in the wandb key, and then it redownloads 5GB of model data. This can be very slow (even with 100mbps...

crispy-cat

feat(docker): Update Dockerfile 🐳 to use `tensorflow:latest-gpu` image

4

The initial base image used often causes a lot of errors between python versions, pip version, JAX versions and CUDA versions as made evident by #260, #234 and #233. This...

SauravMaheshkar

How to generate reproducible results

5

I want to be able to generate the same set of images each time I run the inference script locally on my machine for reproducibility for the same set of...

mandanraf

Suggestion: Switch to Jupyter Labs (No supporting changes needed)

1

Dunno if anyone else noticed but Jupyter has a new thing -- Jupyter Labs. It's a much better interface overall. Zero compatibility issues. Much nicer -- seriously, much much more...

codefaux

Solves #260 by changing the Dockerfile to use a specified version of jaxlib

10

Specifies the exact version of jaxlib if it returns the following error: ``` #5 12.13 ERROR: No matching distribution found for jaxlib==0.3.10+cuda11.cudnn82 (from jax[cuda]) ------ executor failed running [/bin/sh -c...

0xhayleydev

in Google Colab Pro+, flax.errors.ScopeParamShapeError: Inconsistent shap es between value and initializer for parameter "embedding" in "/model/encoder/em bed_tokens"

2

I tried to run the notebook in Colab Pro+, and got: ``` UnfilteredStackTrace Traceback (most recent call last) [](https://localhost:8080/#) in () 19 temperature, ---> 20 cond_scale, 21 ) 46 frames...

dmd

Support offloading encode, for generate() with much less VRAM

6

generate() from Transformers can take encoder outputs as kwargs instead of running the encoder. This PR extends this to "super conditioning" sampling. It also enables providing only one "null sequence"...

drdaxxy

Input your own parameters in the demo

4

https://huggingface.co/spaces/dalle-mini/dalle-mini In the demo inference site, It'd be nice to have a feature that is to be able to input your own parameters such as `seed`, `gen_top_k`, `gen_top_p`, for example...

YukiSakuma

Script stuck when running the script on local device(Macbook Intel)

2

Hi basically when I am running locally by following [inference pipeline notebook](https://github.com/borisdayma/dalle-mini/blob/main/tools/inference/inference_pipeline.ipynb) The script gets stuck at the end when it's generating an image at 0%.

aryan1107