Summary:
Test Plan:
Test that this iterpolates in the right order, config -> configs -> cli args
```
# All three sources
python -m bytelatent.print_config config=bytelatent/configs/debug.yaml configs=[internal/configs/s3_debug.yaml] eval=null
# What worked before
python -m bytelatent.print_config config=internal/configs/s3_debug.yaml eval=null
```
Summary:
- Make the data/checkpoint code fsspec compatible
- Still will not work with s3 saves, due to `torch.distributed.checkpoint.save` not being out of the box workable with `fsspec`. Will implement in followup PR
Test Plan:
Run unit tests and the commands below
```
python -m bytelatent.train config=internal/configs/s3_debug.yaml eval=null checkpoint.dump.every=100
```
```
torchrun --nproc-per-node 8 -m bytelatent.train config=internal/configs/s3_debug.yaml eval=null checkpoint.dump.every=100
```
These currently won't work due to the torch distributed save, but theses hould be tested at a later date
```
python -m bytelatent.train config=internal/configs/s3_debug.yaml eval=null checkpoint.dump.every=100 dump_dir=s3://blt/scratch/checkpoint-test/
```
```
torchrun --nproc-per-node 8 -m bytelatent.train config=internal/configs/s3_debug.yaml eval=null checkpoint.dump.every=100 dump_dir=s3://blt/scratch/checkpoint-test/
```
Summary:
With >1 GPU, but only 1 node, all reduces fail when inputs are not bf16. This uses a modified copy of torch's grad norm to avoid failures
Test Plan:
- Run unit tests:
- Run single gpu training: `python -m bytelatent.train config=internal/configs/s3_debug.yaml eval=null checkpoint.dump.every=100`
- Run 1 node, multi-gpu training `torchrun --nproc-per-node 8 -m bytelatent.train config=internal/configs/s3_debug.yaml eval=null checkpoint.dump.every=100`
Summary:
- Refactor local model configs to be separate and clearer
- Add attention arguments and correct which attention is used in local models
- Preparation for being able to have an entropy train script
- Fix failing unit tests
Test Plan:
