Upload folder using huggingface_hub
Browse files
ema_model/config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_class_name": "MaskGiTUViT_v2",
|
| 3 |
+
"_version": "0.0.1",
|
| 4 |
+
"add_cond_embeds": true,
|
| 5 |
+
"add_micro_cond_embeds": true,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"block_num_heads": 12,
|
| 8 |
+
"block_out_channels": [
|
| 9 |
+
768
|
| 10 |
+
],
|
| 11 |
+
"codebook_size": 8192,
|
| 12 |
+
"cond_embed_dim": 768,
|
| 13 |
+
"decay": 0.9999,
|
| 14 |
+
"encoder_hidden_size": 768,
|
| 15 |
+
"force_down_up_sample": true,
|
| 16 |
+
"hidden_dropout": 0.0,
|
| 17 |
+
"hidden_size": 1024,
|
| 18 |
+
"in_channels": 768,
|
| 19 |
+
"intermediate_size": 2816,
|
| 20 |
+
"inv_gamma": 1.0,
|
| 21 |
+
"layer_norm_eps": 1e-06,
|
| 22 |
+
"ln_elementwise_affine": true,
|
| 23 |
+
"mask_token_id": 8255,
|
| 24 |
+
"micro_cond_embed_dim": 1280,
|
| 25 |
+
"micro_cond_encode_dim": 256,
|
| 26 |
+
"min_decay": 0.0,
|
| 27 |
+
"norm_type": "rmsnorm",
|
| 28 |
+
"num_attention_heads": 16,
|
| 29 |
+
"num_hidden_layers": 22,
|
| 30 |
+
"num_res_blocks": 3,
|
| 31 |
+
"optimization_step": 3000,
|
| 32 |
+
"power": 0.6666666666666666,
|
| 33 |
+
"update_after_step": 0,
|
| 34 |
+
"use_bias": false,
|
| 35 |
+
"use_ema_warmup": false,
|
| 36 |
+
"use_fused_mlp": false,
|
| 37 |
+
"use_fused_residual_norm": true,
|
| 38 |
+
"vocab_size": 8256
|
| 39 |
+
}
|
ema_model/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d86a75a6dcb282dd253d0b36f07cdd4c70d946e180c268e6c6888a912375146a
|
| 3 |
+
size 2433247453
|
text_encoder_lora/README.md
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
### Framework versions
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
- PEFT 0.5.0
|
text_encoder_lora/adapter_config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": {
|
| 3 |
+
"base_model_class": "CLIPTextModelWithProjection",
|
| 4 |
+
"parent_library": "transformers.models.clip.modeling_clip"
|
| 5 |
+
},
|
| 6 |
+
"base_model_name_or_path": "openMUSE/clip-vit-large-patch14-text-enc",
|
| 7 |
+
"bias": "none",
|
| 8 |
+
"fan_in_fan_out": false,
|
| 9 |
+
"inference_mode": true,
|
| 10 |
+
"init_lora_weights": true,
|
| 11 |
+
"layers_pattern": null,
|
| 12 |
+
"layers_to_transform": null,
|
| 13 |
+
"lora_alpha": 32,
|
| 14 |
+
"lora_dropout": 0.0,
|
| 15 |
+
"modules_to_save": null,
|
| 16 |
+
"peft_type": "LORA",
|
| 17 |
+
"r": 128,
|
| 18 |
+
"revision": null,
|
| 19 |
+
"target_modules": [
|
| 20 |
+
"q_proj",
|
| 21 |
+
"k_proj",
|
| 22 |
+
"v_proj"
|
| 23 |
+
],
|
| 24 |
+
"task_type": null
|
| 25 |
+
}
|
text_encoder_lora/adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9c87c9686ef190e071022cf6f12059f1916e013c13baa36fc41590857a9d30d3
|
| 3 |
+
size 28338185
|