Llama 2 13B — RL-MPQ High Fidelity
Standalone RL-MPQ (Reinforcement Learning Mixed-Precision Quantization) checkpoint for the High Fidelity scenario — a quantized variant of meta-llama/Llama-2-13b-hf.
| Field | Value |
|---|---|
| Base model | meta-llama/Llama-2-13b-hf |
| Scenario | High Fidelity |
| Avg bits / weight | 6.7 |
| Compression vs FP16 | 2.3881× |
| WikiText-2 PPL | 4.4313 |
| Layers | 40 |
| Bit distribution | {'4': 13, '8': 27} |
| Format | Fake-quant FP16 + rlmpq_policy.json |
Collection: RL-MPQ — Llama 2 13B — all five scenarios for Llama 2 13B.
Usage
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "AvoCahDoe/llama-2-13b-rlmpq-high-fidelity"
model = AutoModelForCausalLM.from_pretrained(repo, torch_dtype="float16")
tokenizer = AutoTokenizer.from_pretrained(repo)
Other Llama 2 13B scenarios
| Scenario | Avg bits | Compression | WikiText-2 PPL |
|---|---|---|---|
| Aggressive | 3.75 | 4.2667x | 4.5724 |
| Balanced | 4.4 | 3.6364x | 4.4797 |
| Conservative | 5.2 | 3.0769x | 4.4663 |
| Extreme Survival | 2.775 | 5.7658x | 6.1148 |
Grouped archive (all scenarios in one repo): AvoCahDoe/llama-2-13b-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
@misc{rlmpq_llama_2_13b_high-fidelity_2026,
title = {RL-MPQ High Fidelity: Llama 2 13B Mixed-Precision Quantization},
author = {AvoCahDoe},
year = {2026},
url = {https://huggingface.co/AvoCahDoe/llama-2-13b-rlmpq-high-fidelity}
}