Metadata-Version: 2.5
Name: polygramme-cloud
Version: 0.1.1
Summary: Polygramme Cloud client: train LoRA adapters on a hosted trainer, chat with them, download the weights. No torch, no GPU deps.
Project-URL: Homepage, https://polygramme.com
Project-URL: Console, https://polygramme.com/console
Project-URL: Documentation, https://docs.polygramme.com/guides/cloud
Project-URL: Source, https://github.com/polygramme/polygramme-cloud
Author: Polygramme
License-Expression: Apache-2.0
Keywords: fine-tuning,lora,polygramme,polygramme-cloud,rl,rlcli,tinker
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.10
Requires-Dist: click>=8.1
Requires-Dist: httpx>=0.27
Requires-Dist: jinja2>=3
Requires-Dist: tinker==0.25.0
Provides-Extra: dev
Requires-Dist: pytest>=8; extra == 'dev'
Description-Content-Type: text/markdown

# polygramme-cloud

The Polygramme Cloud client. Train LoRA adapters on a hosted trainer, chat with
them the moment a checkpoint lands, and download the weights — from any laptop.
No torch, no GPU dependencies.

```bash
pip install polygramme-cloud
export POLYGRAMME_API_KEY=pg_…          # https://polygramme.com/console → API keys

polygramme train --dataset conversations.jsonl --save-every 25
polygramme chat --model <model_id> "hi"
polygramme checkpoints
polygramme download checkpoint/<model_id>/final.tar.gz
polygramme export <model_id>              # merged full HF checkpoint
```

`conversations.jsonl` is one `{"messages": [{"role": ..., "content": ...}, ...]}`
per line (what `rlcli import` produces). The shared trainer is Qwen3.5-9B with
rank-32 LoRA adapters.

Python:

```python
from polygramme_cloud import Cloud
c = Cloud()                                   # POLYGRAMME_API_KEY from env
for delta in c.chat("<model_id>", [{"role": "user", "content": "hi"}], stream=True):
    print(delta, end="")
```

For training on your own GPUs, use [rlcli](https://github.com/polygramme/rlcli),
the open framework this cloud runs.
