dealignai/GLM-5.3-Flash-UNCENSORED-NVFP4
docs: add measured video fps engine-kill + media parameter warnings
docs: add reasoning_effort / max_tokens serving warning to top of card
fix: default clear_thinking=true in chat_template.jinja + add repetition_penalty=1.1 default in generation_config.json — mitigates known GLM-5.3 reasoning-loop degeneration on multi-turn agent workloads (SGLang #36669, opencode #45533). Upstream Z.ai chat template default `clear_thinking=false` replays prior reasoning_content into subsequent turns, priming a decode attractor. Z.ai's own docs recommend `clear_thinking=true` for chat scenarios — this bakes that in.
fix: default clear_thinking=true in chat_template.jinja + add repetition_penalty=1.1 default in generation_config.json — mitigates known GLM-5.3 reasoning-loop degeneration on multi-turn agent workloads (SGLang #36669, opencode #45533). Upstream Z.ai chat template default `clear_thinking=false` replays prior reasoning_content into subsequent turns, priming a decode attractor. Z.ai's own docs recommend `clear_thinking=true` for chat scenarios — this bakes that in.
Update card: reasoning-mode transparency (off/max fully uncensored, greedy worst-case), refreshed scores
2026-08-28 fixed nvfp4 shard model-input-scales.safetensors
2026-08-28 fixed nvfp4 shard model-00106-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00103-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00101-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00098-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00095-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00092-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00087-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00084-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00081-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00079-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00076-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00073-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00070-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00068-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00065-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00062-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00057-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00056-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00054-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00051-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00048-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00045-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00043-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00040-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00037-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00034-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00032-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00029-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00026-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00023-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00020-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00018-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00015-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00012-of-00120.safetensors
2026-08-28 fixed nvfp4 shard model-00009-of-00120.safetensors
2026-08-28 multimodal template
Release