Upload transformer/config.json with huggingface_hub
Browse files- transformer/config.json +66 -0
transformer/config.json
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_class_name": "PixArtTransformer2DModel",
|
| 3 |
+
"_commit_hash": null,
|
| 4 |
+
"_diffusers_version": "0.32.2",
|
| 5 |
+
"_name_or_path": "/home/ubuntu/.cache/huggingface/hub/models--Jingya--pixart_sigma_pipe_xl_2_512_ms/snapshots/ee9e1f15aa5e778132655fab637943a6bc71a27c/transformer",
|
| 6 |
+
"_use_default_values": [
|
| 7 |
+
"use_additional_conditions"
|
| 8 |
+
],
|
| 9 |
+
"activation_fn": "gelu-approximate",
|
| 10 |
+
"attention_bias": true,
|
| 11 |
+
"attention_head_dim": 72,
|
| 12 |
+
"attention_type": "default",
|
| 13 |
+
"caption_channels": 4096,
|
| 14 |
+
"cross_attention_dim": 1152,
|
| 15 |
+
"double_self_attention": false,
|
| 16 |
+
"dropout": 0.0,
|
| 17 |
+
"in_channels": 4,
|
| 18 |
+
"interpolation_scale": 1,
|
| 19 |
+
"neuron": {
|
| 20 |
+
"auto_cast": "none",
|
| 21 |
+
"auto_cast_type": "bf16",
|
| 22 |
+
"compiler_type": "neuronx-cc",
|
| 23 |
+
"compiler_version": "2.16.372.0+4a9b2326",
|
| 24 |
+
"dynamic_batch_size": false,
|
| 25 |
+
"inline_weights_to_neff": true,
|
| 26 |
+
"input_names": [
|
| 27 |
+
"sample",
|
| 28 |
+
"encoder_hidden_states",
|
| 29 |
+
"timestep",
|
| 30 |
+
"encoder_attention_mask"
|
| 31 |
+
],
|
| 32 |
+
"model_type": "pixart-transformer-2d",
|
| 33 |
+
"optlevel": "2",
|
| 34 |
+
"output_attentions": false,
|
| 35 |
+
"output_hidden_states": false,
|
| 36 |
+
"output_names": [
|
| 37 |
+
"out_hidden_states"
|
| 38 |
+
],
|
| 39 |
+
"static_batch_size": 1,
|
| 40 |
+
"static_encoder_hidden_size": 4096,
|
| 41 |
+
"static_height": 64,
|
| 42 |
+
"static_num_channels": 4,
|
| 43 |
+
"static_patch_size": 2,
|
| 44 |
+
"static_sequence_length": 120,
|
| 45 |
+
"static_vae_scale_factor": 8,
|
| 46 |
+
"static_width": 64,
|
| 47 |
+
"tensor_parallel_size": 1
|
| 48 |
+
},
|
| 49 |
+
"norm_elementwise_affine": false,
|
| 50 |
+
"norm_eps": 1e-06,
|
| 51 |
+
"norm_num_groups": 32,
|
| 52 |
+
"norm_type": "ada_norm_single",
|
| 53 |
+
"num_attention_heads": 16,
|
| 54 |
+
"num_embeds_ada_norm": 1000,
|
| 55 |
+
"num_layers": 28,
|
| 56 |
+
"num_vector_embeds": null,
|
| 57 |
+
"only_cross_attention": false,
|
| 58 |
+
"out_channels": 8,
|
| 59 |
+
"patch_size": 2,
|
| 60 |
+
"sample_size": 64,
|
| 61 |
+
"task": "semantic-segmentation",
|
| 62 |
+
"transformers_version": null,
|
| 63 |
+
"upcast_attention": false,
|
| 64 |
+
"use_additional_conditions": null,
|
| 65 |
+
"use_linear_projection": false
|
| 66 |
+
}
|