Downloads · 30 days
21
2% of all-time downloads
greghavens/fabletron-nemotron-3-super-120b-BF16
fabletron-nemotron-3-super-120b-BF16 is a text generation model from greghavens. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as other.
QLoRA fine-tune of NVIDIA Nemotron-3-Super-120B-A12B (120B-total / 12B-active hybrid Mamba-2 + Latent-MoE, nemotronh) on the piagent split of Glint-Research/Fable-5-traces, targeting reasoning, agentic planning, and t…
Downloads · 30 days
21
2% of all-time downloads
All-time downloads
1.2K
Public
Parameters
124B
247 GB on disk
Likes
4
Public
Click a slice to open those files.
.safetensors247 GB · 100%
How the weights are stored.
BF16124B · 100%
From the Hugging Face model README
QLoRA fine-tune of NVIDIA Nemotron-3-Super-120B-A12B (120B-total / 12B-active hybrid
Mamba-2 + Latent-MoE, nemotron_h) on the pi_agent split of
Glint-Research/Fable-5-traces,
targeting reasoning, agentic planning, and tool-use. This repo holds the full-precision
BF16 weights (the LoRA delta merged into the base, ≈231 GB across 50 shards), loadable directly
with 🤗 Transformers. It is the reference master from which the GGUF / NVFP4 / MLX exports derive.
LM Studio download note: there is a known LM Studio bug where its HF-proxy downloader hangs at "Finalizing download…" on empty (0-byte) files and never marks the model ready — tracked upstream at lmstudio-ai/lmstudio-bug-tracker#1477. These are full-precision Transformers weights, not an LM Studio format (use the MLX-4bit or GGUF siblings for LM Studio) — but the empty
__init__.pythat ships withnemotron_hrepos triggers that bug for anyone pulling this repo through LM Studio's downloader. As a precaution it now ships a non-empty__init__.pyso such downloads finalize normally.
greghavens/fabletron-nemotron-3-super-120b-GGUFgreghavens/fabletron-nemotron-3-super-120b-NVFP4greghavens/fabletron-nemotron-3-super-120b-MLX-4bitimport torch
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "greghavens/fabletron-nemotron-3-super-120b-BF16"
tok = AutoTokenizer.from_pretrained(repo, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
repo, trust_remote_code=True, torch_dtype=torch.bfloat16,
attn_implementation="eager", device_map="auto")
trust_remote_code=Trueandattn_implementation="eager"are required bynemotron_h.- Needs the Mamba-2 kernels (
causal_conv1d,mamba_ssm) for the fast path.- BF16 weights are ≈231 GB → multi-GPU or CPU offload. For one GPU, use a quantized variant above.
- Prompt format is ChatML (
<|im_start|>/<|im_end|>).
q/k/v/o_proj, Mamba-2 in_proj/out_proj, and the
up/down_proj of all 512 routed experts + the shared expert — this MoE is non-gated,
so up/down_proj is the full expert FFN (there is no gate_proj). The MoE router
(mixer.gate, an nn.Parameter) and the MoE-latent fc1/fc2_latent_proj stay frozen.Glint-Research/Fable-5-traces, config pi_agent, ChatML, response-only loss.pi_agent, 81 rows, base → fine-tuned): cross-entropy
1.12 → 0.78, perplexity 3.07 → 2.18.greghavens/fabletron-nemotron-3-super-120b.This is a derivative of two upstream works, and downstream use must comply with both:
Glint-Research/Fable-5-traces — AGPL-3.0, and distilled from Anthropic
Claude outputs (subject to Anthropic's usage terms).Released as a research artifact. Review the upstream licenses before redistribution or
production use. license: other reflects the combination of the above, not a single license.
Research on hybrid Mamba-2/MoE reasoning models, agentic/tool-use experimentation. Not safety-aligned beyond the base model; evaluate before any production deployment.
<!-- FABLETRON_BENCHMARKS_START -->Fabletron is a Fable-5 fine-tune of Nemotron-3-Super-120B-A12B. This repository hosts the BF16 (full-precision safetensors) build. The results below were produced through a deployed NeMo Evaluator at full context length (262 144) with a 32 768-token generation budget, comparing base Nemotron-3-Super against the Fabletron fine-tune in its two served formats — GGUF (llama.cpp) and NVFP4 (vLLM / Blackwell).

| Benchmark | Base Nemotron 3 Super | Fabletron (GGUF) | Fabletron (NVFP4) |
|---|---|---|---|
| IFEval | 89.6 | 90.4 | 87.4 |
| GSM8K | 96.6 | 96.5 | 96.5 |
| MMLU-Pro | 81.1 | 81.5 | 81.2 |
| HumanEval+ | 87.8 | 89.6 | 89.6 |
| MBPP+ | 95.5 | 96.6 | 96.8 |
| BFCL (v3 · multi_turn + ast) | 31.7 | 53.6 | 53.9 |
Scores are % (higher is better). The standout is function-calling (BFCL): 31.7 → 53.6 — base Nemotron largely fails multi-turn and parallel tool use, which the fine-tune resolves (single-tool calls 0.8 → 85.7, multi-turn 0 → 45). General reasoning and coding hold or improve slightly, and NVFP4 tracks the GGUF fine-tune within noise — the only give-back to quantization is IFEval strict (≈2 points).

Mean per-request decode rate across the 6 benchmark families (tok/s, client-side, parallelism 4): base 45.7 · Fabletron GGUF 46.3 · Fabletron NVFP4 63.8. The NVFP4 build serves at ≈1.4× the GGUF (llama.cpp) rate on an RTX PRO 6000 (Blackwell).
<sub>BFCL = Berkeley Function-Calling Leaderboard v3 (multi_turn,ast, native tool-calling, via nvidia_bfcl 26.3). Base and GGUF share the llama.cpp engine; NVFP4 runs on vLLM. Output truncation at this budget is ≈0 across families (MMLU-Pro shows ≈3.5% length-capped generations, intrinsic to the benchmark and near-identical for all three models, so it does not bias the comparison).</sub>