Text-to-Image
Diffusers
Safetensors
Ideogram4Pipeline
sdnq
ideogram
ideogram_4
4-bit precision
8-bit precision
Instructions to use Disty0/Ideogram-4-SDNQ-4bit-dynamic-hadamard with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use Disty0/Ideogram-4-SDNQ-4bit-dynamic-hadamard with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("Disty0/Ideogram-4-SDNQ-4bit-dynamic-hadamard", torch_dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
| { | |
| "add_skip_keys": true, | |
| "dequantize_fp32": true, | |
| "dynamic_loss_threshold": -1, | |
| "group_size": 0, | |
| "hadamard_group_size": 128, | |
| "is_integer": true, | |
| "is_training": false, | |
| "modules_dtype_dict": { | |
| "float5_e3m1fn": [ | |
| "layers.3.adaln_modulation.weight", | |
| "layers.33.adaln_modulation.weight", | |
| "layers.32.adaln_modulation.weight", | |
| "layers.9.adaln_modulation.weight", | |
| "layers.27.adaln_modulation.weight", | |
| "layers.21.adaln_modulation.weight", | |
| "layers.22.adaln_modulation.weight", | |
| "layers.23.adaln_modulation.weight", | |
| "layers.31.adaln_modulation.weight", | |
| "layers.17.adaln_modulation.weight", | |
| "layers.11.adaln_modulation.weight", | |
| "layers.12.adaln_modulation.weight", | |
| "layers.28.adaln_modulation.weight", | |
| "layers.18.adaln_modulation.weight", | |
| "layers.2.adaln_modulation.weight", | |
| "layers.5.adaln_modulation.weight", | |
| "layers.13.adaln_modulation.weight", | |
| "layers.29.adaln_modulation.weight", | |
| "layers.6.adaln_modulation.weight", | |
| "layers.15.adaln_modulation.weight", | |
| "layers.10.adaln_modulation.weight", | |
| "layers.19.adaln_modulation.weight", | |
| "layers.14.adaln_modulation.weight", | |
| "layers.24.adaln_modulation.weight", | |
| "layers.20.adaln_modulation.weight", | |
| "layers.7.adaln_modulation.weight", | |
| "layers.1.adaln_modulation.weight", | |
| "layers.4.adaln_modulation.weight", | |
| "layers.16.adaln_modulation.weight", | |
| "layers.25.adaln_modulation.weight", | |
| "layers.26.adaln_modulation.weight", | |
| "layers.30.adaln_modulation.weight", | |
| "layers.8.adaln_modulation.weight" | |
| ], | |
| "uint4": [ | |
| "layers.12.attention.to_k.weight", | |
| "layers.0.attention.to_k.weight", | |
| "layers.23.attention.to_out.0.weight", | |
| "layers.18.attention.to_k.weight", | |
| "layers.3.attention.to_k.weight", | |
| "layers.16.attention.to_out.0.weight", | |
| "layers.12.attention.to_q.weight", | |
| "layers.0.attention.to_out.0.weight", | |
| "layers.28.attention.to_out.0.weight", | |
| "layers.33.attention.to_k.weight", | |
| "layers.23.attention.to_q.weight", | |
| "layers.18.attention.to_q.weight", | |
| "layers.19.feed_forward.w3.weight", | |
| "layers.3.attention.to_q.weight", | |
| "layers.16.attention.to_q.weight", | |
| "layers.8.feed_forward.w2.weight", | |
| "layers.12.attention.to_v.weight", | |
| "layers.0.attention.to_q.weight", | |
| "layers.33.attention.to_q.weight", | |
| "layers.21.feed_forward.w1.weight", | |
| "layers.23.attention.to_v.weight", | |
| "layers.18.attention.to_v.weight", | |
| "layers.2.attention.to_out.0.weight", | |
| "layers.3.attention.to_v.weight", | |
| "layers.16.attention.to_v.weight", | |
| "layers.13.attention.to_k.weight", | |
| "layers.0.attention.to_v.weight", | |
| "layers.10.feed_forward.w2.weight", | |
| "layers.33.attention.to_v.weight", | |
| "layers.24.feed_forward.w3.weight", | |
| "layers.9.attention.to_out.0.weight", | |
| "layers.20.feed_forward.w1.weight", | |
| "layers.19.attention.to_out.0.weight", | |
| "layers.28.feed_forward.w1.weight", | |
| "layers.6.feed_forward.w2.weight", | |
| "layers.13.attention.to_q.weight", | |
| "layers.32.feed_forward.w1.weight", | |
| "layers.27.attention.to_out.0.weight", | |
| "layers.21.attention.to_k.weight", | |
| "layers.13.attention.to_v.weight", | |
| "layers.21.feed_forward.w3.weight", | |
| "layers.23.feed_forward.w2.weight", | |
| "layers.2.feed_forward.w1.weight", | |
| "layers.3.feed_forward.w3.weight", | |
| "layers.16.feed_forward.w2.weight", | |
| "layers.10.feed_forward.w3.weight", | |
| "layers.21.attention.to_q.weight", | |
| "layers.33.feed_forward.w3.weight", | |
| "layers.0.feed_forward.w3.weight", | |
| "layers.15.attention.to_out.0.weight", | |
| "layers.9.feed_forward.w2.weight", | |
| "layers.16.attention.to_k.weight", | |
| "layers.8.feed_forward.w1.weight", | |
| "layers.21.attention.to_v.weight", | |
| "layers.2.feed_forward.w2.weight", | |
| "layers.23.attention.to_k.weight", | |
| "layers.27.feed_forward.w1.weight", | |
| "layers.28.feed_forward.w2.weight", | |
| "layers.23.feed_forward.w3.weight", | |
| "layers.19.feed_forward.w1.weight", | |
| "layers.6.feed_forward.w1.weight", | |
| "layers.32.feed_forward.w3.weight", | |
| "layers.17.attention.to_out.0.weight", | |
| "layers.31.feed_forward.w2.weight", | |
| "layers.1.feed_forward.w1.weight", | |
| "layers.22.attention.to_v.weight", | |
| "layers.4.attention.to_v.weight", | |
| "layers.24.attention.to_q.weight", | |
| "layers.6.attention.to_q.weight", | |
| "layers.26.attention.to_k.weight", | |
| "layers.15.attention.to_v.weight", | |
| "layers.17.attention.to_q.weight", | |
| "layers.19.attention.to_q.weight", | |
| "layers.24.attention.to_v.weight", | |
| "layers.6.attention.to_v.weight", | |
| "layers.26.attention.to_q.weight", | |
| "layers.17.attention.to_v.weight", | |
| "layers.14.feed_forward.w2.weight", | |
| "layers.19.attention.to_v.weight", | |
| "layers.28.attention.to_k.weight", | |
| "layers.0.feed_forward.w1.weight", | |
| "layers.25.attention.to_k.weight", | |
| "layers.7.attention.to_k.weight", | |
| "layers.30.feed_forward.w3.weight", | |
| "layers.11.attention.to_k.weight", | |
| "layers.26.attention.to_v.weight", | |
| "layers.4.feed_forward.w3.weight", | |
| "layers.22.feed_forward.w1.weight", | |
| "layers.12.feed_forward.w2.weight", | |
| "layers.28.attention.to_q.weight", | |
| "layers.1.attention.to_k.weight", | |
| "layers.18.attention.to_out.0.weight", | |
| "layers.25.attention.to_q.weight", | |
| "layers.15.feed_forward.w2.weight", | |
| "layers.2.attention.to_k.weight", | |
| "layers.7.attention.to_out.0.weight", | |
| "layers.31.attention.to_out.0.weight", | |
| "layers.20.feed_forward.w3.weight", | |
| "layers.7.feed_forward.w3.weight", | |
| "layers.11.attention.to_q.weight", | |
| "layers.28.attention.to_v.weight", | |
| "layers.1.attention.to_q.weight", | |
| "layers.25.attention.to_v.weight", | |
| "layers.5.attention.to_out.0.weight", | |
| "layers.2.attention.to_q.weight", | |
| "layers.7.attention.to_q.weight", | |
| "layers.9.attention.to_k.weight", | |
| "layers.22.attention.to_k.weight", | |
| "layers.32.attention.to_out.0.weight", | |
| "layers.11.attention.to_v.weight", | |
| "layers.14.feed_forward.w3.weight", | |
| "layers.1.attention.to_v.weight", | |
| "layers.26.feed_forward.w1.weight", | |
| "layers.2.attention.to_v.weight", | |
| "layers.9.attention.to_q.weight", | |
| "layers.7.attention.to_v.weight", | |
| "layers.4.attention.to_k.weight", | |
| "layers.22.attention.to_out.0.weight", | |
| "layers.22.feed_forward.w2.weight", | |
| "layers.18.feed_forward.w2.weight", | |
| "layers.6.attention.to_k.weight", | |
| "layers.15.attention.to_k.weight", | |
| "layers.13.feed_forward.w1.weight", | |
| "layers.16.feed_forward.w3.weight", | |
| "layers.19.attention.to_k.weight", | |
| "layers.20.feed_forward.w2.weight", | |
| "layers.3.feed_forward.w2.weight", | |
| "layers.15.attention.to_q.weight", | |
| "layers.11.feed_forward.w2.weight", | |
| "layers.4.attention.to_q.weight", | |
| "layers.25.feed_forward.w1.weight", | |
| "layers.26.feed_forward.w2.weight", | |
| "layers.17.attention.to_k.weight", | |
| "layers.22.attention.to_q.weight", | |
| "layers.9.attention.to_v.weight", | |
| "layers.6.attention.to_out.0.weight", | |
| "layers.7.feed_forward.w2.weight", | |
| "layers.24.attention.to_k.weight", | |
| "layers.10.attention.to_out.0.weight", | |
| "layers.10.feed_forward.w1.weight", | |
| "layers.10.attention.to_q.weight", | |
| "layers.30.attention.to_k.weight", | |
| "layers.10.attention.to_v.weight", | |
| "layers.24.attention.to_out.0.weight", | |
| "layers.30.attention.to_q.weight", | |
| "layers.21.feed_forward.w2.weight", | |
| "layers.19.feed_forward.w2.weight", | |
| "layers.8.feed_forward.w3.weight", | |
| "layers.6.feed_forward.w3.weight", | |
| "layers.12.attention.to_out.0.weight", | |
| "layers.27.attention.to_k.weight", | |
| "layers.0.feed_forward.w2.weight", | |
| "layers.14.attention.to_k.weight", | |
| "layers.27.feed_forward.w3.weight", | |
| "layers.33.feed_forward.w2.weight", | |
| "layers.30.attention.to_v.weight", | |
| "layers.27.attention.to_q.weight", | |
| "layers.14.attention.to_out.0.weight", | |
| "layers.29.attention.to_out.0.weight", | |
| "layers.20.attention.to_k.weight", | |
| "layers.15.feed_forward.w1.weight", | |
| "layers.31.feed_forward.w3.weight", | |
| "layers.8.attention.to_k.weight", | |
| "layers.1.attention.to_out.0.weight", | |
| "layers.24.feed_forward.w2.weight", | |
| "layers.4.attention.to_out.0.weight", | |
| "layers.27.attention.to_v.weight", | |
| "layers.14.attention.to_q.weight", | |
| "layers.20.attention.to_q.weight", | |
| "layers.22.feed_forward.w3.weight", | |
| "layers.33.attention.to_out.0.weight", | |
| "layers.8.attention.to_out.0.weight", | |
| "layers.9.feed_forward.w1.weight", | |
| "layers.12.feed_forward.w1.weight", | |
| "layers.14.attention.to_v.weight", | |
| "layers.30.feed_forward.w1.weight", | |
| "layers.20.attention.to_v.weight", | |
| "layers.8.attention.to_q.weight", | |
| "layers.26.attention.to_out.0.weight", | |
| "layers.13.attention.to_out.0.weight", | |
| "layers.29.feed_forward.w3.weight", | |
| "layers.33.feed_forward.w1.weight", | |
| "layers.31.feed_forward.w1.weight", | |
| "layers.13.feed_forward.w3.weight", | |
| "layers.14.feed_forward.w1.weight", | |
| "layers.21.attention.to_out.0.weight", | |
| "layers.9.feed_forward.w3.weight", | |
| "layers.27.feed_forward.w2.weight", | |
| "layers.17.feed_forward.w2.weight", | |
| "layers.23.feed_forward.w1.weight", | |
| "layers.26.feed_forward.w3.weight", | |
| "layers.8.attention.to_v.weight", | |
| "layers.5.feed_forward.w3.weight", | |
| "layers.3.attention.to_out.0.weight", | |
| "layers.1.feed_forward.w2.weight", | |
| "layers.10.attention.to_k.weight", | |
| "layers.32.attention.to_q.weight", | |
| "layers.5.attention.to_q.weight", | |
| "layers.20.attention.to_out.0.weight", | |
| "layers.32.attention.to_v.weight", | |
| "layers.5.attention.to_v.weight", | |
| "layers.31.attention.to_k.weight", | |
| "layers.1.feed_forward.w3.weight", | |
| "layers.3.feed_forward.w1.weight", | |
| "layers.29.feed_forward.w1.weight", | |
| "layers.18.feed_forward.w3.weight", | |
| "layers.17.feed_forward.w3.weight", | |
| "layers.16.feed_forward.w1.weight", | |
| "layers.13.feed_forward.w2.weight", | |
| "layers.31.attention.to_q.weight", | |
| "layers.11.feed_forward.w3.weight", | |
| "layers.29.attention.to_k.weight", | |
| "layers.11.attention.to_out.0.weight", | |
| "layers.30.attention.to_out.0.weight", | |
| "layers.24.feed_forward.w1.weight", | |
| "layers.31.attention.to_v.weight", | |
| "layers.29.attention.to_q.weight", | |
| "layers.4.feed_forward.w2.weight", | |
| "layers.32.feed_forward.w2.weight", | |
| "layers.7.feed_forward.w1.weight", | |
| "layers.5.feed_forward.w1.weight", | |
| "layers.25.feed_forward.w3.weight", | |
| "layers.25.attention.to_out.0.weight", | |
| "layers.15.feed_forward.w3.weight", | |
| "layers.12.feed_forward.w3.weight", | |
| "layers.17.feed_forward.w1.weight", | |
| "layers.32.attention.to_k.weight", | |
| "layers.28.feed_forward.w3.weight", | |
| "layers.30.feed_forward.w2.weight", | |
| "layers.25.feed_forward.w2.weight", | |
| "layers.2.feed_forward.w3.weight", | |
| "layers.5.feed_forward.w2.weight", | |
| "layers.4.feed_forward.w1.weight", | |
| "layers.18.feed_forward.w1.weight", | |
| "layers.29.feed_forward.w2.weight", | |
| "layers.29.attention.to_v.weight", | |
| "layers.5.attention.to_k.weight", | |
| "layers.11.feed_forward.w1.weight" | |
| ] | |
| }, | |
| "modules_quant_config": {}, | |
| "modules_to_not_convert": [ | |
| "layers.0.adaln_modulation.weight", | |
| "t_embedding", | |
| "llm_cond_proj", | |
| "final_layer", | |
| "llm_cond_norm", | |
| "input_proj", | |
| "embed_image_indicator", | |
| "adaln_proj", | |
| "layers.0.attention.norm_q.weight", | |
| "layers.11.ffn_norm1.weight", | |
| "layers.12.attention.norm_q.weight", | |
| "layers.28.ffn_norm1.weight", | |
| "layers.29.attention_norm1.weight", | |
| "layers.17.attention_norm1.weight", | |
| "layers.18.attention.norm_k.weight", | |
| "layers.18.attention.norm_q.weight", | |
| "layers.3.attention.norm_k.weight", | |
| "layers.3.attention.norm_q.weight", | |
| "layers.32.ffn_norm2.weight", | |
| "layers.33.attention.norm_q.weight", | |
| "layers.10.attention_norm2.weight", | |
| "layers.24.attention.norm_q.weight", | |
| "layers.2.ffn_norm1.weight", | |
| "layers.2.ffn_norm2.weight", | |
| "layers.28.attention_norm1.weight", | |
| "layers.28.attention_norm2.weight", | |
| "layers.31.ffn_norm1.weight", | |
| "layers.32.attention_norm2.weight", | |
| "layers.2.attention.norm_k.weight", | |
| "layers.12.attention_norm2.weight", | |
| "layers.12.ffn_norm2.weight", | |
| "layers.9.attention.norm_k.weight", | |
| "layers.23.attention_norm1.weight", | |
| "layers.18.attention_norm1.weight", | |
| "layers.18.attention_norm2.weight", | |
| "layers.2.attention_norm1.weight", | |
| "layers.2.attention_norm2.weight", | |
| "layers.16.attention_norm1.weight", | |
| "layers.0.attention_norm1.weight", | |
| "layers.0.attention_norm2.weight", | |
| "layers.33.attention_norm1.weight", | |
| "layers.24.ffn_norm1.weight", | |
| "layers.25.ffn_norm2.weight", | |
| "layers.26.attention_norm1.weight", | |
| "layers.26.ffn_norm1.weight", | |
| "layers.26.ffn_norm2.weight", | |
| "layers.9.attention_norm1.weight", | |
| "layers.9.attention_norm2.weight", | |
| "layers.19.attention_norm1.weight", | |
| "layers.27.attention.norm_k.weight", | |
| "layers.21.attention.norm_k.weight", | |
| "layers.7.ffn_norm1.weight", | |
| "layers.7.ffn_norm2.weight", | |
| "layers.27.attention_norm2.weight", | |
| "layers.14.attention.norm_q.weight", | |
| "layers.14.ffn_norm2.weight", | |
| "layers.22.attention_norm2.weight", | |
| "layers.22.ffn_norm1.weight", | |
| "layers.30.ffn_norm1.weight", | |
| "layers.16.ffn_norm2.weight", | |
| "layers.23.attention.norm_q.weight", | |
| "layers.10.ffn_norm2.weight", | |
| "layers.31.attention.norm_k.weight", | |
| "layers.33.ffn_norm2.weight", | |
| "layers.4.attention_norm1.weight", | |
| "layers.4.attention_norm2.weight", | |
| "layers.5.attention.norm_k.weight", | |
| "layers.5.attention_norm1.weight", | |
| "layers.5.ffn_norm2.weight", | |
| "layers.6.attention.norm_q.weight", | |
| "layers.6.attention_norm1.weight", | |
| "layers.6.attention_norm2.weight", | |
| "layers.0.ffn_norm1.weight", | |
| "layers.1.attention_norm1.weight", | |
| "layers.1.attention_norm2.weight", | |
| "layers.17.attention.norm_k.weight", | |
| "layers.9.ffn_norm1.weight", | |
| "layers.14.attention_norm1.weight", | |
| "layers.14.attention_norm2.weight", | |
| "layers.26.attention.norm_k.weight", | |
| "layers.3.ffn_norm1.weight", | |
| "layers.30.attention.norm_q.weight", | |
| "layers.30.attention_norm1.weight", | |
| "layers.30.attention_norm2.weight", | |
| "layers.11.ffn_norm2.weight", | |
| "layers.22.attention_norm1.weight", | |
| "layers.12.attention.norm_k.weight", | |
| "layers.12.attention_norm1.weight", | |
| "layers.27.attention.norm_q.weight", | |
| "layers.15.attention_norm1.weight", | |
| "layers.24.attention_norm1.weight", | |
| "layers.24.ffn_norm2.weight", | |
| "layers.25.attention.norm_k.weight", | |
| "layers.6.ffn_norm1.weight", | |
| "layers.6.ffn_norm2.weight", | |
| "layers.7.attention.norm_k.weight", | |
| "layers.7.attention.norm_q.weight", | |
| "layers.10.ffn_norm1.weight", | |
| "layers.17.attention_norm2.weight", | |
| "layers.17.ffn_norm2.weight", | |
| "layers.19.attention_norm2.weight", | |
| "layers.19.ffn_norm1.weight", | |
| "layers.19.ffn_norm2.weight", | |
| "layers.0.ffn_norm2.weight", | |
| "layers.1.attention.norm_k.weight", | |
| "layers.2.attention.norm_q.weight", | |
| "layers.4.ffn_norm2.weight", | |
| "layers.15.ffn_norm1.weight", | |
| "layers.13.attention.norm_k.weight", | |
| "layers.13.attention_norm1.weight", | |
| "layers.13.attention_norm2.weight", | |
| "layers.31.attention_norm1.weight", | |
| "layers.32.attention.norm_q.weight", | |
| "layers.20.ffn_norm1.weight", | |
| "layers.20.ffn_norm2.weight", | |
| "layers.21.attention_norm2.weight", | |
| "layers.21.ffn_norm1.weight", | |
| "layers.21.ffn_norm2.weight", | |
| "layers.8.attention.norm_q.weight", | |
| "layers.8.attention_norm2.weight", | |
| "layers.5.attention_norm2.weight", | |
| "layers.29.attention.norm_k.weight", | |
| "layers.3.attention_norm1.weight", | |
| "layers.4.attention.norm_k.weight", | |
| "layers.11.attention_norm1.weight", | |
| "layers.14.ffn_norm1.weight", | |
| "layers.15.attention.norm_k.weight", | |
| "layers.1.ffn_norm2.weight", | |
| "layers.20.attention.norm_q.weight", | |
| "layers.20.attention_norm1.weight", | |
| "layers.20.attention_norm2.weight", | |
| "layers.7.attention_norm2.weight", | |
| "layers.22.ffn_norm2.weight", | |
| "layers.23.attention.norm_k.weight", | |
| "layers.23.ffn_norm2.weight", | |
| "layers.10.attention.norm_k.weight", | |
| "layers.13.ffn_norm1.weight", | |
| "layers.13.ffn_norm2.weight", | |
| "layers.19.attention.norm_q.weight", | |
| "layers.21.attention_norm1.weight", | |
| "layers.8.attention_norm1.weight", | |
| "layers.17.ffn_norm1.weight", | |
| "layers.18.ffn_norm2.weight", | |
| "layers.19.attention.norm_k.weight", | |
| "layers.0.attention.norm_k.weight", | |
| "layers.3.attention_norm2.weight", | |
| "layers.9.ffn_norm2.weight", | |
| "layers.24.attention.norm_k.weight", | |
| "layers.15.attention.norm_q.weight", | |
| "layers.15.attention_norm2.weight", | |
| "layers.10.attention_norm1.weight", | |
| "layers.11.attention.norm_k.weight", | |
| "layers.11.attention_norm2.weight", | |
| "layers.24.attention_norm2.weight", | |
| "layers.14.attention.norm_k.weight", | |
| "layers.27.ffn_norm1.weight", | |
| "layers.27.ffn_norm2.weight", | |
| "layers.28.attention.norm_k.weight", | |
| "layers.28.ffn_norm2.weight", | |
| "layers.7.attention_norm1.weight", | |
| "layers.33.ffn_norm1.weight", | |
| "layers.1.attention.norm_q.weight", | |
| "layers.32.ffn_norm1.weight", | |
| "layers.33.attention.norm_k.weight", | |
| "layers.16.attention.norm_q.weight", | |
| "layers.16.ffn_norm1.weight", | |
| "layers.4.ffn_norm1.weight", | |
| "layers.27.attention_norm1.weight", | |
| "layers.25.attention.norm_q.weight", | |
| "layers.25.attention_norm2.weight", | |
| "layers.23.attention_norm2.weight", | |
| "layers.33.attention_norm2.weight", | |
| "layers.26.attention.norm_q.weight", | |
| "layers.12.ffn_norm1.weight", | |
| "layers.13.attention.norm_q.weight", | |
| "layers.31.attention.norm_q.weight", | |
| "layers.31.attention_norm2.weight", | |
| "layers.29.ffn_norm1.weight", | |
| "layers.29.ffn_norm2.weight", | |
| "layers.29.attention_norm2.weight", | |
| "layers.20.attention.norm_k.weight", | |
| "layers.15.ffn_norm2.weight", | |
| "layers.16.attention.norm_k.weight", | |
| "layers.16.attention_norm2.weight", | |
| "layers.30.ffn_norm2.weight", | |
| "layers.21.attention.norm_q.weight", | |
| "layers.22.attention.norm_k.weight", | |
| "layers.22.attention.norm_q.weight", | |
| "layers.23.ffn_norm1.weight", | |
| "layers.32.attention_norm1.weight", | |
| "layers.5.ffn_norm1.weight", | |
| "layers.6.attention.norm_k.weight", | |
| "layers.29.attention.norm_q.weight", | |
| "layers.1.ffn_norm1.weight", | |
| "layers.10.attention.norm_q.weight", | |
| "layers.11.attention.norm_q.weight", | |
| "layers.3.ffn_norm2.weight", | |
| "layers.30.attention.norm_k.weight", | |
| "layers.18.ffn_norm1.weight", | |
| "layers.17.attention.norm_q.weight", | |
| "layers.31.ffn_norm2.weight", | |
| "layers.32.attention.norm_k.weight", | |
| "layers.5.attention.norm_q.weight", | |
| "layers.4.attention.norm_q.weight", | |
| "layers.25.ffn_norm1.weight", | |
| "layers.26.attention_norm2.weight", | |
| "layers.28.attention.norm_q.weight", | |
| "layers.25.attention_norm1.weight", | |
| "layers.8.attention.norm_k.weight", | |
| "layers.8.ffn_norm1.weight", | |
| "layers.8.ffn_norm2.weight", | |
| "layers.9.attention.norm_q.weight" | |
| ], | |
| "modules_to_not_use_matmul": [], | |
| "non_blocking": false, | |
| "quant_conv": false, | |
| "quant_embedding": false, | |
| "quant_method": "sdnq", | |
| "quantization_device": "cuda", | |
| "quantized_matmul_dtype": null, | |
| "return_device": "cpu", | |
| "sdnq_version": "0.2.0", | |
| "svd_rank": 32, | |
| "svd_steps": 8, | |
| "use_dynamic_quantization": true, | |
| "use_grad_ckpt": true, | |
| "use_hadamard": true, | |
| "use_quantized_matmul": false, | |
| "use_quantized_matmul_conv": false, | |
| "use_static_quantization": true, | |
| "use_stochastic_rounding": false, | |
| "use_svd": false, | |
| "weights_dtype": "uint4" | |
| } | |