swcrazyfan
swcrazyfan
Added preliminary TensorFlow (and MacBook M1 GPU) support for text generation by creating two new classes: TFHappyTransformer and TFHappyGeneration. However, since I created two full new classes by copying and...
When I turn on FP16, only the steps show. Because of this, I have no idea how well the fine-tuning is going and whether I need to keep going or...
**Describe the solution you'd like** I'd love to be able to fine-tune the style/grammar of the resulting sentences without needing to have sentence and keyword pairs--only sentences. I'm experimenting with...
Currently, aitextgen does not work to train anything larger than the 125M model with GPT-NEO (the 350M model is no longer available). However, the official GPT-NEO Colab notebook uses TPU...
I'm experimenting with fine-tuning, but I've recently realized loss can decrease, but the model overfits. Therefore, it can actually become not as good of a model. Do you have a...
Is it possible to implement the ability to train GPT-NEO from scratch as opposed to only GPT2?
I'm trying to train a model based on GPT Neo 125M, and I keep getting this error. It continues to train and even create text, but I'm pretty sure this...
If I want to set a static learning rate, would I write it here? `--scale_lr` Or is it somewhere else? I'd like to experiment with higher/lower learning rate and higher/lower...
I've been running this on Paperspace for a while and it always does the "cache latents" step, but recently that disappeared. I tried using my old diffusers folder, and it...
Up until today, your dreambooth script would first cache latents. Then, train. Up until today, I had no issues. Today, it's just not doing anything no matter if I reinstall...