Qwen3.8-Flash-Next Uncensored, 4-bit codebook experts + int8 trunk — radiance container

orcarouter/Qwen3.8-Flash-Next-Uncensored — Qwen3.8-Flash-Next with its refusal direction abliterated — quantised from its bf16 checkpoint into a single .rad container for the radiance inference engine (AMD RDNA4, ROCm), with the model's MTP head for speculative decoding and its vision tower. It is the same recipe as StillDeadcode/qwen3.8-next-flash-fp8-iq4r-moe.

This model's safety alignment has been removed: it answers requests the original model refuses. The source model's card describes how it was made, how it was evaluated and what it is for.


File	`qwen3.8-next-flash-uncensored-fp8-iq4r-moe.rad` — 113.6 GiB
Routed experts	4-bit codes into a 16-level non-uniform codebook after a 128-point Walsh–Hadamard rotation, one E4M3 scale per 64 weights; codes chosen by GPTQ (10M-token calibration, per-expert down-projection Hessians) and refined by three sweeps of coordinate descent
Protected experts	ten experts that carry most of their layer's down-projection energy, kept bf16
Trunk	attention, delta-net and shared-expert linears and the lm_head int8 (W8A8), a scale per 128 columns searched for least error; hyper-connection mixing matrices E4M3
Speculator	the model's MTP head (depth 3 by default)
Vision	the vision tower, bf16: images and video in chat requests
Context	262,144 tokens trained; served at 200K

Quality

The calibration data was recorded on the original model, and this file was not measured against its own bf16 weights. For the recipe's error, the same recipe on the original model, as full-vocabulary KL divergence against its bf16, teacher-forced over ~75K positions of chat, tool-use and code transcripts:

	mean KL	median	p99	p99.9	top-1 agreement
all tokens	0.0789	0.0037	1.65	5.33	92.85%
text and assistant turns	0.0239	0.0029	0.32	1.03	94.57%
reference: a second bf16 implementation, all tokens	0.0399	0.0013	0.82	3.43	95.13%

Serve

The routed experts do not fit two 32 GB cards; the engine keeps the hottest in VRAM and streams the rest from a pinned host pool, so the host needs about 32 GB of free RAM.

radiance --model qwen3.8-next-flash-uncensored-fp8-iq4r-moe.rad --tp 2 --tp-wire wht6 \
    --max-model-len 200000 --max-num-seqs 8 --kv-cache-dtype fp8 \
    --placement expert_tiered --host-pool-mib 12288 --gpu-headroom-mib 96 \
    --expert-vs-cache-ratio 0.82 --num-speculative-tokens 3 --max-num-batched-tokens 2048 \
    --host 0.0.0.0 --port 8000

The recipe is served on 2× Radeon AI PRO R9700 (gfx1201). This file was converted there with no errors and uploaded without being served first.

The server speaks the OpenAI API (/v1/chat/completions, /v1/completions), with tool calls, structured output, and image_url / video parts in chat messages.

How this file was made

CALIB=calib/w4nl-calib rad-convert orcarouter/Qwen3.8-Flash-Next-Uncensored \
    --tokenizer orcarouter/Qwen3.8-Flash-Next-Uncensored/tokenizer.json \
    --recipe qwen4exp-w4nl64-i8-hc8m.recipe -o qwen3.8-next-flash-uncensored-fp8-iq4r-moe.rad

The recipe (qwen4exp-w4nl64-i8-hc8m.recipe in this repository; $CALIB is its calibration data, which is not published). The first rule that matches a weight decides it, and everything no rule names is the checkpoint's bf16:

blk.34.ffn_*_exps.407.weight cast dtype=bf16
blk.34.ffn_*_exps.496.weight cast dtype=bf16
blk.44.ffn_*_exps.292.weight cast dtype=bf16
blk.44.ffn_*_exps.350.weight cast dtype=bf16
blk.46.ffn_*_exps.290.weight cast dtype=bf16
blk.46.ffn_*_exps.392.weight cast dtype=bf16
blk.47.ffn_*_exps.122.weight cast dtype=bf16
blk.47.ffn_*_exps.143.weight cast dtype=bf16
blk.47.ffn_*_exps.399.weight cast dtype=bf16
blk.47.ffn_*_exps.445.weight cast dtype=bf16
blk.*.ffn_*_exps.*.weight gptq table=w4nl group=64 scale=fp8_e4m3 scale2=f32 block2=*x* scale2_value=0.0001220703125 transform=fwht128 rule=search cd=3 calib=$CALIB
*_hc_down.weight rtn codes=fp8_e4m3 group=128 scale=f32
*_hc_up.weight rtn codes=fp8_e4m3 group=80 scale=f32
mtp.fc_hidden.weight rtn codes=fp8_e4m3 group=128 scale=f32
mtp.fc_embedding.weight rtn codes=fp8_e4m3 group=128 scale=f32
blk.*.attn_qg.weight rtn codes=i8 group=128 scale=bf16 clamp=sym rule=search
blk.*.attn_k.weight rtn codes=i8 group=128 scale=bf16 clamp=sym rule=search
blk.*.attn_v.weight rtn codes=i8 group=128 scale=bf16 clamp=sym rule=search
blk.*.attn_output.weight rtn codes=i8 group=128 scale=bf16 clamp=sym rule=search
blk.*.ssm_inz.weight rtn codes=i8 group=128 scale=bf16 clamp=sym rule=search
blk.*.ssm_out.weight rtn codes=i8 group=128 scale=bf16 clamp=sym rule=search
blk.*.ffn_gate_up_shexp.weight rtn codes=i8 group=128 scale=bf16 clamp=sym rule=search
blk.*.ffn_down_shexp.weight rtn codes=i8 group=128 scale=bf16 clamp=sym rule=search
output.weight rtn codes=i8 group=128 scale=bf16 clamp=sym rule=search
blk.*.ple_ngram.weight rtn codes=fp8_e4m3 block=*x* scale=bf16 rule=fixed scale_value=1.99317932128906e-4
mtp.draft_head.weight rtn codes=u2 zero=u8 group=128 scale=f16

License

The Qwen community license of the base model; see LICENSE. The source model's card states Apache-2.0, but its repository ships this license file, the one the base model carries.