Anique

Bug: Loss redefined twice for different variables Fix: The second loss variable is renamed

```python3 def denoising_score_matching(scorenet, samples, key, sigma=0.01): noise = jax.random.normal(key, samples.shape) perturbed_samples = samples + noise * sigma target = -noise / sigma scores = scorenet(perturbed_samples) loss = 1 / 2....

In `prepare_data.py` I found the following code: ``` train_lable_distribution.append([pre_vote_image_name, en_im_path, ld]) ``` Here `pre_vote_image_name` and `ld` are not changed inside the loop. So its essentially adding the same thing over...

Thanks for the effort to port Llama 2 to jax. After looking further in the code, it seems that the config for the 13B model is missing. Relevant file: ```...

the variable `VARIANT` is defined capital but was used in lowercase `variant`