
                    For jax, we pick up the best solutions provided by github life savers.
                    If google brings you here, you are so lucky! These life savers are angels, they have well solved your problems.
                    So do not waste time scrolling the original issue threads just to dig out solutions... 
                    Here you are, you are welcome. Then, go back to work!
                    

I managed to solve the issue by installing Rust compiler

Install Rust link curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
Restart the terminal
pip install transformers==2.5.1

transformers Installation Error - Failed building wheel for tokenizers

You should have obtained the 12 layers as well as the embedding output ❓ Questions & Help
I tried to set the output_hidden_states=True 

transformers How to get all layers(12) hidden states of BERT?

change
base_output = base_model([ids mask token_type_ids])
to
base_output = base_model.bert([ids mask token_type_ids])
should fix 

transformers Failing to load saved TFBertModel

This means you're encoding a sequence that is larger than the max sequence the model can handle (which is 512 tokens) 

transformers token indices sequence length is longer than the specified maximum sequence length

Hi! I have tried to get jaxlib working now on my apple m1 Hi all I was digging around to see what might need to happen to allow JAX to work on Apple Silicon 

jax Provide wheels for macOS ARM

I had this problem when I trained the model with torch==1.6.0 and tried to load the model with 1.3.1 

transformers Unable to load weights from pytorch checkpoint file

Update for latest transformers add modeling_bert.py:78:

and convert_bert_original_tf_checkpoint_to_pytorch.py 

transformers Bert for passage reranking

Hi two datasets are available in run_lm_finetuning.py:

TextDataset ❓ Questions & Help
Details
I wanted to employ the examples/run_lm_finetuning.py from the transformers repository on a pretrained Bert model

transformers Input file format for examples/run_lm_finetuning.py

Same issue 🐛 Bug
Information
I am trying to build a Keras Sequential model where I use DistillBERT as a non-trainable embedding layer 

transformers model.save() does not save keras model that includes DIstillBert layer

It should be possible to do something like this for now:

And then you should be able to use it with the LineByLineTextDataset 

transformers TypeError: 'ByteLevelBPETokenizer' object is not callable

Dig into this a little When computing sentence probability do we need to prepend the sentence with a dummy start token (e.g 

transformers GPT2 Sentence Probability: Necessary to Prepend "<|endoftext|>"?

No need to downgrade the transformers Just do the following - it's from the migration guide. 

transformers dropout(): argument 'input' (position 1) must be Tensor, not str With Bert

@WilliamNurmi @knuser :
Fine-tuning the albert_xxlargev1_sqd2_512_bs48 script with Transformers 2.4.1 yielded the following results:


which is no improvement over fine-tuning the same script with Transformers 2.3.0
My best model to date is now posted at:  https://huggingface.co/ahotrod/albert_xxlargev1_squad2_512
You can access this albert_xxlargev1_sqd2_512 fine-tuned model with:


The AutoModels: (AutoConfig

transformers XLNET SQuAD2.0 Fine-Tuning - What May Have Changed?

I found a solution The model.generate() function is necessary to extract the predictions 

transformers How to use fine-tuned BART for prediction?

@jotaf98 it is done and available in master branch Axes of size 1 can be written as 1 not () 

einops [Feature suggestion] Identifiers not on both sides of the expression

I managed to quantize Pytorch GPT-2 XL to int8 with quantize_torch_model method from this script 

transformers GPT-2 XL PyTorch Quantization for use on a Cloud Server

The pip install -e  is probably working it's just that some tests are failing due to code not tests on Torch 1.2.0 The install should have worked fine 