hackit/mate-local-models
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Add local models via Git LFS (23 models, ~143 GB)

Browse source
This commit is contained in:
hackit 2026-06-28 17:25:21 +00:00
parent 94b5c12c2d
commit 9b381fb3de
246 changed files with 38788 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

27
image/mlx/prism-ml/bonsai-image-ternary-4B-mlx-2bit/transformer-packed-mflux/config.json Normal file
View file

@ -0,0 +1,27 @@
{
"_class_name": "Flux2Transformer2DModel",
"_diffusers_version": "0.37.1",
"_name_or_path": "black-forest-labs/FLUX.2-klein-4B",
"attention_head_dim": 128,
"axes_dims_rope": [
32,
32,
32,
32
],
"enable_time_sign_embed": false,
"eps": 1e-06,
"guidance_embeds": false,
"in_channels": 128,
"joint_attention_dim": 7680,
"mlp_ratio": 3.0,
"musubi_block_swap_device": "cpu",
"musubi_blocks_to_swap": 0,
"num_attention_heads": 24,
"num_layers": 5,
"num_single_layers": 20,
"out_channels": null,
"patch_size": 1,
"rope_theta": 2000,
"timestep_guidance_channels": 256
}

BIN
image/mlx/prism-ml/bonsai-image-ternary-4B-mlx-2bit/transformer-packed-mflux/diffusion_pytorch_model.safetensors (Stored with Git LFS) Normal file
View file

Binary file not shown.

120
image/mlx/prism-ml/bonsai-image-ternary-4B-mlx-2bit/transformer-packed-mflux/quantization_config.json Normal file
View file

@ -0,0 +1,120 @@
{
"format": "mlx-packed-affine",
"solver": "ternary",
"bits": 2,
"group_size": 128,
"scale_dtype": "bfloat16",
"skip_patterns": [
"proj_out",
"x_embedder",
"context_embedder",
"time_text_embed",
"time_guidance_embed",
"norm_out",
"double_stream_modulation_img",
"double_stream_modulation_txt",
"single_stream_modulation"
],
"quantized_modules": [
"single_transformer_blocks.0.attn.to_out",
"single_transformer_blocks.0.attn.to_qkv_mlp_proj",
"single_transformer_blocks.1.attn.to_out",
"single_transformer_blocks.1.attn.to_qkv_mlp_proj",
"single_transformer_blocks.10.attn.to_out",
"single_transformer_blocks.10.attn.to_qkv_mlp_proj",
"single_transformer_blocks.11.attn.to_out",
"single_transformer_blocks.11.attn.to_qkv_mlp_proj",
"single_transformer_blocks.12.attn.to_out",
"single_transformer_blocks.12.attn.to_qkv_mlp_proj",
"single_transformer_blocks.13.attn.to_out",
"single_transformer_blocks.13.attn.to_qkv_mlp_proj",
"single_transformer_blocks.14.attn.to_out",
"single_transformer_blocks.14.attn.to_qkv_mlp_proj",
"single_transformer_blocks.15.attn.to_out",
"single_transformer_blocks.15.attn.to_qkv_mlp_proj",
"single_transformer_blocks.16.attn.to_out",
"single_transformer_blocks.16.attn.to_qkv_mlp_proj",
"single_transformer_blocks.17.attn.to_out",
"single_transformer_blocks.17.attn.to_qkv_mlp_proj",
"single_transformer_blocks.18.attn.to_out",
"single_transformer_blocks.18.attn.to_qkv_mlp_proj",
"single_transformer_blocks.19.attn.to_out",
"single_transformer_blocks.19.attn.to_qkv_mlp_proj",
"single_transformer_blocks.2.attn.to_out",
"single_transformer_blocks.2.attn.to_qkv_mlp_proj",
"single_transformer_blocks.3.attn.to_out",
"single_transformer_blocks.3.attn.to_qkv_mlp_proj",
"single_transformer_blocks.4.attn.to_out",
"single_transformer_blocks.4.attn.to_qkv_mlp_proj",
"single_transformer_blocks.5.attn.to_out",
"single_transformer_blocks.5.attn.to_qkv_mlp_proj",
"single_transformer_blocks.6.attn.to_out",
"single_transformer_blocks.6.attn.to_qkv_mlp_proj",
"single_transformer_blocks.7.attn.to_out",
"single_transformer_blocks.7.attn.to_qkv_mlp_proj",
"single_transformer_blocks.8.attn.to_out",
"single_transformer_blocks.8.attn.to_qkv_mlp_proj",
"single_transformer_blocks.9.attn.to_out",
"single_transformer_blocks.9.attn.to_qkv_mlp_proj",
"transformer_blocks.0.attn.add_k_proj",
"transformer_blocks.0.attn.add_q_proj",
"transformer_blocks.0.attn.add_v_proj",
"transformer_blocks.0.attn.to_add_out",
"transformer_blocks.0.attn.to_k",
"transformer_blocks.0.attn.to_out.0",
"transformer_blocks.0.attn.to_q",
"transformer_blocks.0.attn.to_v",
"transformer_blocks.0.ff.linear_in",
"transformer_blocks.0.ff.linear_out",
"transformer_blocks.0.ff_context.linear_in",
"transformer_blocks.0.ff_context.linear_out",
"transformer_blocks.1.attn.add_k_proj",
"transformer_blocks.1.attn.add_q_proj",
"transformer_blocks.1.attn.add_v_proj",
"transformer_blocks.1.attn.to_add_out",
"transformer_blocks.1.attn.to_k",
"transformer_blocks.1.attn.to_out.0",
"transformer_blocks.1.attn.to_q",
"transformer_blocks.1.attn.to_v",
"transformer_blocks.1.ff.linear_in",
"transformer_blocks.1.ff.linear_out",
"transformer_blocks.1.ff_context.linear_in",
"transformer_blocks.1.ff_context.linear_out",
"transformer_blocks.2.attn.add_k_proj",
"transformer_blocks.2.attn.add_q_proj",
"transformer_blocks.2.attn.add_v_proj",
"transformer_blocks.2.attn.to_add_out",
"transformer_blocks.2.attn.to_k",
"transformer_blocks.2.attn.to_out.0",
"transformer_blocks.2.attn.to_q",
"transformer_blocks.2.attn.to_v",
"transformer_blocks.2.ff.linear_in",
"transformer_blocks.2.ff.linear_out",
"transformer_blocks.2.ff_context.linear_in",
"transformer_blocks.2.ff_context.linear_out",
"transformer_blocks.3.attn.add_k_proj",
"transformer_blocks.3.attn.add_q_proj",
"transformer_blocks.3.attn.add_v_proj",
"transformer_blocks.3.attn.to_add_out",
"transformer_blocks.3.attn.to_k",
"transformer_blocks.3.attn.to_out.0",
"transformer_blocks.3.attn.to_q",
"transformer_blocks.3.attn.to_v",
"transformer_blocks.3.ff.linear_in",
"transformer_blocks.3.ff.linear_out",
"transformer_blocks.3.ff_context.linear_in",
"transformer_blocks.3.ff_context.linear_out",
"transformer_blocks.4.attn.add_k_proj",
"transformer_blocks.4.attn.add_q_proj",
"transformer_blocks.4.attn.add_v_proj",
"transformer_blocks.4.attn.to_add_out",
"transformer_blocks.4.attn.to_k",
"transformer_blocks.4.attn.to_out.0",
"transformer_blocks.4.attn.to_q",
"transformer_blocks.4.attn.to_v",
"transformer_blocks.4.ff.linear_in",
"transformer_blocks.4.ff.linear_out",
"transformer_blocks.4.ff_context.linear_in",
"transformer_blocks.4.ff_context.linear_out"
]
}