Merge branch 'lab/dashboard' into 'export/v1-2-0'

add dashboard

See merge request onecomp/onecomp-lab!77

Author: aki916f

.gitignore

@@ -29,3 +29,29 @@ tests/onecomp/quantizer/jointq/data/model_layers_0_self_attn_k_proj.pth
 .uv-sync-vllm.lock
 .uv-sync-vllm.done
 .venv-vllm/
+
+# ── dashboard/ (HPC LLM quantization dashboard) ──────────────
+# Re-include the app/models package shadowed by the global `models` rule above
+!dashboard/backend/app/models/
+!dashboard/backend/app/models/*.py
+
+# uv venv (.venv-vllm/ already covered above)
+dashboard/backend/.venv/
+
+# Redis built from source per dashboard/docs/setup-hpc.md §1.3
+dashboard/backend/7.2.7.tar.gz
+dashboard/backend/redis-*/
+dashboard/backend/dump.rdb
+
+# Runtime state (SQLite, quantized models)
+dashboard/backend/onecomp.db
+dashboard/backend/onecomp.db-journal
+dashboard/backend/tmp/
+
+# Frontend (node_modules already in dashboard/frontend/.gitignore; dist/ covered by global rule)
+dashboard/frontend/node_modules/
+# yarn is canonical (README uses `yarn dev`); avoid double lockfiles
+dashboard/frontend/package-lock.json
+
+# vLLM / worker logs written under /tmp by setup-hpc.md
+*.log

CHANGELOG.md

@@ -3,6 +3,12 @@
 
 ## [v1.2.0] 2026-06-04 (WIP)
 
+## New Feature : Dashboard
+
+- Added `dashboard/`, a browser-based web app for OneCompression on **SLURM-managed HPC GPU nodes without Docker**: pick a Hugging Face model and quantization settings in the UI, run jobs on the GPU, deploy the quantized checkpoint, and validate inference via chat
+- **Stack**: React + Vite frontend (local PC), FastAPI API, Celery worker + user-built Redis, SQLite job DB, per-job output under `backend/tmp/quantized/`; CUDA quantization via `onecomp` and chat deploy via a separate **vLLM** subprocess from the same `backend/.venv` (`onecomp` + `vllm>=0.21` in `pyproject.toml`)
+- **Quantization methods exposed in the UI**: `gptq`, `autobit`, `jointq`, and `auto_run` (VRAM-based bitwidth / group size); optional QEP (not with JointQ); fractional bit widths for `autobit` / `auto_run`
+
 ### for Developer: pre-commit
 
 - Added `.pre-commit-config.yaml` with `black`, `isort`, and local hooks (`no-japanese`, `copyright-header`, `no-email-address`); install with `uv sync --extra dev` then `pre-commit install` (see README)

README.md

@@ -41,6 +41,7 @@ Full documentation is available at **[https://FujitsuResearch.github.io/OneCompr
 - **Block-wise PTQ**: Post-quantization block-wise distillation that minimises intermediate-representation MSE against an FP16 teacher model at Transformer-block granularity. Includes Phase 1 (greedy per-block optimisation) and Phase 2 CBQ (cross-block sliding-window optimisation). Supports GPTQ, DBF, and OneBit quantizers.
 - **LoRA SFT Post-Process**: Fine-tune quantized models with LoRA adapters for accuracy recovery or domain-specific knowledge injection. Supports SFT loss, teacher distillation, and intermediate block alignment.
 - **Rotation Preprocessing**: SpinQuant/OstQuant-based rotation preprocessing that reduces quantization error by learning optimal rotation matrices before quantization. Rotation/scaling matrices are absorbed into model weights, with online Hadamard hooks automatically registered at load time. Supports Llama and Qwen3 architectures.
+- **Web Dashboard (HPC)**: A browser-based dashboard for launching quantization jobs, deploying models, and validating chat-based inference in HPC environments. See [dashboard/README.md](dashboard/README.md) for details.
 - (TBD)
 
 ## 🤖 Supported Models