huggingface/tokenizers fork error on startup with gpu

[TheWilhelm]

I downloaded the Repo and requirements yesterday, it did work without the GPU library.
However since i have loaded the relevant GPU library i get a error when trying to start any model.
I tried to set the environment variable but it did not have any effect.

any help is greatly appreciated

[]$ python3.10 app.py --port 8080 --model_version Mega
--> Starting DALL-E Server. This might take up to two minutes.
Some of the weights of DalleBart were initialized in float16 precision from the model checkpoint at /tmp/tmpbz2cg4_7:
[('lm_head', 'kernel'), ('model', 'decoder', 'embed_positions', 'embedding'), ('model', 'decoder', 'embed_tokens', 'embedding'), ('model', 'decoder', 'final_ln', 'bias'), ('model', 'decoder', 'layernorm_embedding', 'bias'), ('model', 'decoder', 'layernorm_embedding', 'scale'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'FlaxBartAttention_0', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'FlaxBartAttention_0', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'FlaxBartAttention_0', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'FlaxBartAttention_0', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'FlaxBartAttention_1', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'FlaxBartAttention_1', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'FlaxBartAttention_1', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'FlaxBartAttention_1', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'GLU_0', 'Dense_0', 'kernel'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'GLU_0', 'Dense_1', 'kernel'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'GLU_0', 'Dense_2', 'kernel'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'GLU_0', 'LayerNorm_0', 'bias'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'GLU_0', 'LayerNorm_1', 'bias'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'LayerNorm_0', 'bias'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'LayerNorm_1', 'bias'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'LayerNorm_1', 'scale'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'LayerNorm_2', 'bias'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'LayerNorm_3', 'bias'), ('model', 'decoder', 'layers', 'FlaxBartDecoderLayers', 'LayerNorm_3', 'scale'), ('model', 'encoder', 'embed_positions', 'embedding'), ('model', 'encoder', 'embed_tokens', 'embedding'), ('model', 'encoder', 'final_ln', 'bias'), ('model', 'encoder', 'layernorm_embedding', 'bias'), ('model', 'encoder', 'layernorm_embedding', 'scale'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'FlaxBartAttention_0', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'FlaxBartAttention_0', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'FlaxBartAttention_0', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'FlaxBartAttention_0', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'GLU_0', 'Dense_0', 'kernel'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'GLU_0', 'Dense_1', 'kernel'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'GLU_0', 'Dense_2', 'kernel'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'GLU_0', 'LayerNorm_0', 'bias'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'GLU_0', 'LayerNorm_1', 'bias'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'LayerNorm_0', 'bias'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'LayerNorm_1', 'bias'), ('model', 'encoder', 'layers', 'FlaxBartEncoderLayers', 'LayerNorm_1', 'scale')]
You should probably UPCAST the model weights to float32 if this was not intended. See [`~FlaxPreTrainedModel.to_fp32`] for further information on how to do this.
huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
To disable this warning, you can either:
        - Avoid using `tokenizers` before the fork if possible
        - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
To disable this warning, you can either:
        - Avoid using `tokenizers` before the fork if possible
        - Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
Aborted (core dumped)
[]$ /usr/local/lib/python3.10/multiprocessing/resource_tracker.py:224: UserWarning: resource_tracker: There appear to be 6 leaked semaphore objects to clean up at shutdown
  warnings.warn('resource_tracker: There appear to be %d '