加载中…
加载中…
This is an ExLlamaV3 quantization at 4.0 bpw (-hq) of Ornith-1.5-35B-A3B-DFlash2 by Resolvo Cloud. The original model card is below.
DFlash2 draft for ornith-ai/Ornith-1.5-35B-A3B-MLX-4bit, produced by the no-op-graft + selector-training recipe:
ornith-ai/Ornith-1.5-35B-A3B-DFlash (DFlashV1) weights, copied verbatim.attention_conv/mlp_conv grafted as identity (unit self-tap, zero dynamic corrections).hidden_projection + predecessor/successor codebooks trained on features captured from the quantized 4-bit target (741 samples, top-16 hard-negative margin, convs frozen).opencode_sim n=4, engine decode)| Draft | serial | packed B4 | speedup |
|---|---|---|---|
| Ornith-1.5-35B-A3B-DFlash (V1) | 142.8 tok/s | 147.4 tok/s | 1.05x |
| jzinno/Ornith-1.5-35B-A3B-DFlash2 (bf16-trained) | 78.3 | 114.6 | 1.32x |
| this checkpoint | 190.1 tok/s | 316.8 tok/s | 1.65x |
The bf16-trained DFlash2 underperforms on the 4-bit target — its selector was calibrated to bf16 hidden states. This checkpoint trained on the deployed quant's own features, which is why it scales with packed lanes.
architectures: ["DFlash2DraftModel"], hidden 2048, dynamic 512, conv_group_size 16, block_size 8, selector top_k 16.Hidden2048 draft context core (draft_conv_h2048).