Upload vae_encoder/config.json with huggingface_hub
Browse files- vae_encoder/config.json +72 -0
vae_encoder/config.json
ADDED
|
@@ -0,0 +1,72 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_class_name": "AutoencoderKL",
|
| 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/vae",
|
| 6 |
+
"_use_default_values": [
|
| 7 |
+
"latents_std",
|
| 8 |
+
"shift_factor",
|
| 9 |
+
"latents_mean",
|
| 10 |
+
"use_quant_conv",
|
| 11 |
+
"mid_block_add_attention",
|
| 12 |
+
"use_post_quant_conv"
|
| 13 |
+
],
|
| 14 |
+
"act_fn": "silu",
|
| 15 |
+
"block_out_channels": [
|
| 16 |
+
128,
|
| 17 |
+
256,
|
| 18 |
+
512,
|
| 19 |
+
512
|
| 20 |
+
],
|
| 21 |
+
"down_block_types": [
|
| 22 |
+
"DownEncoderBlock2D",
|
| 23 |
+
"DownEncoderBlock2D",
|
| 24 |
+
"DownEncoderBlock2D",
|
| 25 |
+
"DownEncoderBlock2D"
|
| 26 |
+
],
|
| 27 |
+
"force_upcast": false,
|
| 28 |
+
"in_channels": 3,
|
| 29 |
+
"latent_channels": 4,
|
| 30 |
+
"latents_mean": null,
|
| 31 |
+
"latents_std": null,
|
| 32 |
+
"layers_per_block": 2,
|
| 33 |
+
"mid_block_add_attention": true,
|
| 34 |
+
"neuron": {
|
| 35 |
+
"auto_cast": "none",
|
| 36 |
+
"auto_cast_type": "bf16",
|
| 37 |
+
"compiler_type": "neuronx-cc",
|
| 38 |
+
"compiler_version": "2.16.372.0+4a9b2326",
|
| 39 |
+
"dynamic_batch_size": false,
|
| 40 |
+
"inline_weights_to_neff": true,
|
| 41 |
+
"input_names": [
|
| 42 |
+
"sample"
|
| 43 |
+
],
|
| 44 |
+
"model_type": "vae-encoder",
|
| 45 |
+
"optlevel": "2",
|
| 46 |
+
"output_attentions": false,
|
| 47 |
+
"output_hidden_states": false,
|
| 48 |
+
"output_names": [
|
| 49 |
+
"latent_parameters"
|
| 50 |
+
],
|
| 51 |
+
"static_batch_size": 1,
|
| 52 |
+
"static_height": 512,
|
| 53 |
+
"static_num_channels": 3,
|
| 54 |
+
"static_width": 512,
|
| 55 |
+
"tensor_parallel_size": 1
|
| 56 |
+
},
|
| 57 |
+
"norm_num_groups": 32,
|
| 58 |
+
"out_channels": 3,
|
| 59 |
+
"sample_size": 512,
|
| 60 |
+
"scaling_factor": 0.13025,
|
| 61 |
+
"shift_factor": null,
|
| 62 |
+
"task": "semantic-segmentation",
|
| 63 |
+
"transformers_version": null,
|
| 64 |
+
"up_block_types": [
|
| 65 |
+
"UpDecoderBlock2D",
|
| 66 |
+
"UpDecoderBlock2D",
|
| 67 |
+
"UpDecoderBlock2D",
|
| 68 |
+
"UpDecoderBlock2D"
|
| 69 |
+
],
|
| 70 |
+
"use_post_quant_conv": true,
|
| 71 |
+
"use_quant_conv": true
|
| 72 |
+
}
|