W4A16-AWQ — 4-bit weight-only (AWQ, compressed-tensors), FP16 activations. Runs on 1Cat-vLLM's SM70 TurboMind path (Tesla V100) and on stock vLLM (Ampere+).

Base model: philbert440/Qwen3.6-40B-DeckardUncensored-OpusDistilled-W4A16-AWQ

Serving

bash

# 1Cat-vLLM on 2x Tesla V100 (SM70): W4A16-AWQ + MTP. (Also runs on stock vLLM, Ampere+.)
export VLLM_SM70_FLASH_ATTN_V100=1 VLLM_SM70_QUANT_BACKEND=turbomind
python -m vllm.entrypoints.openai.api_server \
  --model philbert440/Qwen3.6-40B-DeckardUncensored-OpusDistilled-HermesCalibrated-W4A16-AWQ --trust-remote-code --dtype half \
  --attention-backend FLASH_ATTN_V100 --tensor-parallel-size 2 \
  --kv-cache-dtype fp8_e5m2 --max-num-seqs 6 \
  --speculative-config '{"method":"mtp","num_speculative_tokens":4,"attention_backend":"FLASH_ATTN_V100","draft_sample_method":"greedy"}' \
  --compilation-config '{"cudagraph_mode":"full_and_piecewise","cudagraph_capture_sizes":[1,2,4,8]}'

Variants

See the Deckard-40B Uncensored collection for all tiers and formats (BF16 / W4A16-AWQ / NVFP4).

Abliteration removes safety refusals; you are responsible for lawful, ethical use.

Changelog

2026-08-15 — tokenizer fix. tokenizer.json / tokenizer_config.json were re-serialized by the llm-compressor calibration run and shipped with an active truncation block (max_length 1024/2048) plus a drifted pre-tokenizer regex (and, on the Qwen3.6-based repos, 7 phantom audio/TTS special tokens the base model does not define). That broke image inputs larger than the limit under transformers 5 / vLLM (Mismatch in image token count, surfacing as an HTTP 400 Failed to apply Qwen3VLProcessor). Both files are now byte-identical to the upstream base model's (vocab/merges/added tokens were always identical — this is a metadata-only restore). If you downloaded before this date, re-fetch those two files. Thanks to @elBuffo for the report.