transformers VS Swin-Transformer-Tensorflow

* Superconvergence + LR range finder + Fast AI's Ranger21 optimizer was the goto optimizer for CNNs, and worked fabulously well, but on transformers, the learning rate range finder sadi 1e-3 was the best, whilst 1e-5 was better. However, the 1 cycle learning rate stuck. https://github.com/huggingface/transformers/issues/16013

Thanks! :) I'm pushing them into transformers, pytorch-gemma and collabing with the Gemma team to resolve all the issues :)

The RoPE fix should already be in transformers 4.38.2: https://github.com/huggingface/transformers/pull/29285

My main PR for transformers which fixes most of the issues (some still left): https://github.com/huggingface/transformers/pull/29402

If you want to tinker with the architecture Hugging Face has a FOSS implementation in transformers: https://github.com/huggingface/transformers/blob/main/src/tr...

If you want to reproduce the training pipeline, you couldn't do that even if you wanted to because you don't have access to thousands of A100s.

In transformers, they tried really hard to have a single function or method to deal with both self and cross attention mechanisms, masking, positional and relative encodings, interpolation etc. While it allows a user to use the same function/method for any model, it has led to severe parameter bloat. Just compare the original implementation of llama by FAIR with the implementation by HF to get an idea.

You can train it with the code provided in transformer repo: https://github.com/huggingface/transformers/blob/main/examples/pytorch/language-modeling/run_clm.py

transformers uses accelerate if you call it with device_map='auto'

Too much money being thrown around on BS in the LLM space, hardly any of it is going to places where it matters.

For example, the researchers working hard on better text sampling techniques, or on better constraint techniques (i.e. like this https://arxiv.org/abs/2306.03081), or on actual negative prompting/CFG in LLMs (i.e. like this https://github.com/huggingface/transformers/issues/24536) are doing far FAR more to advance the state of AI than dozens of VC backed LLM "prompt engineering" companies operating today.

HN, and the NLP community have some serious blindspots with knowing how to exploit their own technology. At least someone at Anderson Howartz got a clue and gave some funding to Oogabooga - still waiting for Automatic1111 to get any funding.

Repo : https://github.com/huggingface/transformers

From what I know, Whisper already supports prompting (https://github.com/huggingface/transformers/pull/22496). Can I somehow freeze the whole model and tune exclusively the prompt or would I need to write an implementation from scratch?

https://github.com/huggingface/transformers/blob/0a55d9f7376...

To summarize how they work: you keep some number of previously generated tokens, and once you get logits that you want to sample a new token from, you find the logits for existing tokens and multiply them by a penalty, thus lowering the probability of the corresponding tokens.