Standalone RL-MPQ (Reinforcement Learning Mixed-Precision Quantization) checkpoint for the High Fidelity scenario — a quantized variant of google/gemma-2-9b.

Field	Value
Base model	google/gemma-2-9b
Scenario	High Fidelity
Avg bits / weight	7.0476
Compression vs FP16	2.2703×
WikiText-2 PPL	104.8098
Layers	42
Bit distribution	`{'4': 10, '8': 32}`
Format	Fake-quant FP16 + `rlmpq_policy.json`

Collection: RL-MPQ — Gemma 2 9B — all five scenarios for Gemma 2 9B.

Usage

python

from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "AvoCahDoe/gemma-2-9b-rlmpq-high-fidelity"

model = AutoModelForCausalLM.from_pretrained(repo, torch_dtype="float16")
tokenizer = AutoTokenizer.from_pretrained(repo)

Other Gemma 2 9B scenarios

Scenario	Avg bits	Compression	WikiText-2 PPL
Aggressive	3.6667	4.3636x	162.8437
Balanced	4.2857	3.7333x	127.0798
Conservative	5.1429	3.1111x	116.5244
Extreme Survival	2.7857	5.7436x	424.7991

Grouped archive (all scenarios in one repo): AvoCahDoe/gemma-2-9b-rlmpq

Method

Phase 3 — PPO agent assigns per-layer bit widths under the High Fidelity reward target.
Phase 4 — Policy replayed on real weights; WikiText-2 perplexity validates quality.
Export — Fake-quantized FP16 weights compatible with Hugging Face Transformers.

Files

File	Description
`config.json`	Llama architecture + RL-MPQ metadata
`model.safetensors`	Fake-quantized weights
`rlmpq_policy.json`	Per-layer bit-width policy
`rlmpq_metrics.json`	Validation & PPL summary

Citation

bibtex

@misc{rlmpq_gemma_2_9b_high-fidelity_2026,
  title  = {RL-MPQ High Fidelity: Gemma 2 9B Mixed-Precision Quantization},
  author = {AvoCahDoe},
  year   = {2026},
  url    = {https://huggingface.co/AvoCahDoe/gemma-2-9b-rlmpq-high-fidelity}
}