| # bigscience | |
| [Research workshop on large language models - The Summer of Language Models 21](https://bigscience.huggingface.co/) | |
| At the moment we have 2 code repos: | |
| 1. https://github.com/bigscience-workshop/Megatron-DeepSpeed - this is our flagship code base | |
| 2. https://github.com/bigscience-workshop/bigscience - (this repo) for everything else - docs, experiments, etc. | |
| Currently, the most active segments of this repo are: | |
| - [JZ](./jz/) - Lots of information about our work environment which helps evaluate, plan and get things done | |
| - [Experiments](./experiments) - many experiments are being done. Documentation, result tables, scripts and logs are all there | |
| - [Datasets info](./data/) | |
| - [Train](./train) - all the information about the current trainings (see below for the most important ones) | |
| We have READMEs for specific aspects, such as: | |
| - [hub integration](./tools/README.md) | |
| ## Trainings | |
| While we keep detailed chronicles of experiments and findings for some of the main trainings, here is a doc that contains a summary of the most important findings: [Lessons learned](train/lessons-learned.md) | |
| ### Train 1 - 13B - unmodified Megatron gpt2 - baseline | |
| * [the full spec and discussions](./train/tr1-13B-base) | |
| * [the training script](./train/tr1-13B-base/tr1-13B-round1.slurm) | |
| * checkpoints and logs: | |
| - [tensorboard](https://huggingface.co/bigscience/tr1-13B-tensorboard/tensorboard) | |
| - [logs](https://huggingface.co/bigscience/tr1-13B-logs/) | |
| * [chronicles](./train/tr1-13B-base/chronicles.md) | |
| You can watch the training logs live by running this `tail -f` like script over remote log file that gets synced to the hub once an hour: | |
| ``` | |
| perl -e '$u=shift; $b=0; while(1){($e)=qx[curl -sI $u]=~/content-length: (\d+)/; \ | |
| print qx[curl -sr $b-$e -L $u] if $e>$b; $b=$e; sleep 300}' \ | |
| https://huggingface.co/bigscience/tr1-13B-logs/resolve/main/main_log.txt | |
| ``` | |
| ### Train 3 | |
| Architecture and scaling baseline runs: no fancy tricks, just GPT2. Here are links to the respective tensorboards: | |
| | Size | 1B3 | 760M | 350M | 125M | | |
| |--------------------- |----- |------ |------ |------ | | |
| | C4 + low warmup | [a](https://huggingface.co/bigscience/tr3-1B3-modeling-baseline-tensorboard) | [b](https://huggingface.co/bigscience/tr3b-760M-modeling-baseline-tensorboard) | [c](https://huggingface.co/bigscience/tr3c-350M-modeling-baseline-tensorboard) | | | |
| | OSCAR + low warmup | [f](https://huggingface.co/bigscience/tr3f-1B3-diagnostic2-low-warmup-oscar-tensorboard) | | | | | |
| | C4 + high warmup | [e](https://huggingface.co/bigscience/tr3e-1B3-diagnostic1-warmup-c4-tensorboard) | | | | | |
| | OSCAR + high warmup | **[d (current baseline)](https://huggingface.co/bigscience/tr3d-1B3-more-warmup-tensorboard)** | [g](https://huggingface.co/bigscience/tr3g-760M-v2-tensorboard) | [h](https://huggingface.co/bigscience/tr3h-350M-v2-tensorboard) | [i](https://huggingface.co/bigscience/tr3i-125M-v2-tensorboard) | | |
| | Pile + high warmup | [m](https://huggingface.co/bigscience/tr3m-1B3-pile-tensorboard) | [j](https://huggingface.co/bigscience/tr3j-760M-pile-tensorboard) | [k](https://huggingface.co/bigscience/tr3k-350M-pile-tensorboard) | [l](https://huggingface.co/bigscience/tr3l-125M-pile-tensorboard) | | |
| ### Train 8 | |
| 104B - unmodified Megatron gpt2 - with extra-wide hidden size to learn how to deal with training instabilities | |
| * [the full spec and discussions](./train/tr8-104B-wide) | |
| * [the training script](./train/tr8-104B-wide/tr8-104B.slurm) | |
| * checkpoints and logs: | |
| - [tensorboard](https://huggingface.co/bigscience/tr8-104B-logs/tensorboard) | |
| - [logs](https://huggingface.co/bigscience/tr8-104B-logs/tree/main/logs) | |
| * [chronicles](./train/tr8-104B-wide/chronicles.md) | |
| You can watch the training logs live by running this `tail -f` like script over remote log file that gets synced to the hub once an hour: | |
| ``` | |
| perl -e '$u=shift; $b=0; while(1){($e)=qx[curl -sI $u]=~/content-length: (\d+)/; \ | |
| print qx[curl -sr $b-$e -L $u] if $e>$b; $b=$e; sleep 300}' \ | |
| https://cdn-lfs.huggingface.co/bigscience/tr8-104B-logs/b2cc478d5ae7c9ec937ea2db1d2fe09de593fa2ec38c171d6cc5dca094cd79f9 | |
| ``` | |
| ### Train 11 | |
| **This is the current main training** | |
| tr11-176B-ml | |
| * [the full spec and discussions](./train/tr11-176B-ml/) | |
| * [the training script](./train/tr11-176B-ml/tr11-176B-ml.slurm) | |
| * checkpoints and logs: | |
| - [tensorboard](https://huggingface.co/bigscience/tr11-176B-ml-logs/tensorboard) | |
| - [logs](https://huggingface.co/bigscience/tr11-176B-ml-logs/tree/main/logs/main) | |
| * [chronicles-prequel](./train/tr11-176B-ml/chronicles-prequel.md) | |
| * [chronicles](./train/tr11-176B-ml/chronicles.md) | |
| You can watch the training logs live by running this `tail -f` like script over remote log file that gets synced to the hub once an hour: | |
| ``` | |
| perl -e '$u=shift; $b=0; while(1){($e)=qx[curl -LsI $u]=~/2 200.*?content-length: (\d+)/s; \ | |
| print qx[curl -Lsr $b-$e $u] if $e>$b; $b=$e; sleep 300}' \ | |
| https://huggingface.co/bigscience/tr11-176B-ml-logs/resolve/main/logs/main/main_log.txt | |
| ``` | |