2511

train_vae_fdl.py

@@ -53,20 +53,21 @@ clip_grad_norm     = 1.0
 mixed_precision    = "no"
 gradient_accumulation_steps = 2
 generated_folder   = "samples"
-save_as            = "vae"
+save_as            = "vae4"
 num_workers        = 0
 device = None
 
 # --- Режимы обучения ---
 # QWEN: учим только декодер
 train_decoder_only = True
+train_up_only      = False
 full_training      = False  # если True — учим весь VAE и добавляем KL (ниже)
 kl_ratio           = 0.00
 
 # Доли лоссов
 loss_ratios = {
-    "lpips": 0.50,
-    "fdl" :  0.25,   
+    "lpips": 0.70,#0.50,
+    "fdl" :  0.05,#0.25,   
     "edge":  0.05,
     "mse":   0.10,
     "mae":   0.10,
@@ -171,13 +172,13 @@ if full_training and not train_decoder_only:
 else:
     # учим только 0-й блок декодера + post_quant_conv
     if hasattr(core, "decoder"):
-        if hasattr(core.decoder, "up_blocks") and len(core.decoder.up_blocks) > 0:
+        if train_up_only:#hasattr(core.decoder, "up_blocks") and len(core.decoder.up_blocks) > 0:
             # --- только 0-й up_block ---
             for name, p in core.decoder.up_blocks[0].named_parameters():
                 p.requires_grad = True
                 unfrozen_param_names.append(f"{name}")
         else:
-            print("[WARN] Decoder has no up_blocks — fallback to full decoder")
+            print("Decoder — fallback to full decoder")
             for name, p in core.decoder.named_parameters():
                 p.requires_grad = True
                 unfrozen_param_names.append(f"decoder.{name}")

vae/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:36ae5585635b638056890063a3382d57ab5a3bfbc34f89825191c0e529015f9f
+oid sha256:1a534956925d8f565c3f17704f955bc09ae316b6876d528b0b2e2cc3db723d32
 size 382598708

vae3/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:95458d45a65f10704a9ce7c75176dcead07708273ae647a40acfb8d9ebc54106
+oid sha256:1a534956925d8f565c3f17704f955bc09ae316b6876d528b0b2e2cc3db723d32
 size 382598708

vae4/config.json

@@ -0,0 +1,48 @@
+{
+  "_class_name": "AsymmetricAutoencoderKL",
+  "_diffusers_version": "0.35.2",
+  "_name_or_path": "vae",
+  "act_fn": "silu",
+  "block_out_channels": [
+    128,
+    128,
+    256,
+    512,
+    512
+  ],
+  "down_block_out_channels": [
+    128,
+    256,
+    512,
+    512
+  ],
+  "down_block_types": [
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D",
+    "DownEncoderBlock2D"
+  ],
+  "force_upcast": false,
+  "in_channels": 3,
+  "latent_channels": 16,
+  "layers_per_down_block": 2,
+  "layers_per_up_block": 2,
+  "norm_num_groups": 32,
+  "out_channels": 3,
+  "sample_size": 1024,
+  "scaling_factor": 1.0,
+  "up_block_out_channels": [
+    128,
+    128,
+    256,
+    512,
+    512
+  ],
+  "up_block_types": [
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D",
+    "UpDecoderBlock2D"
+  ]
+}

vae4/diffusion_pytorch_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:98a644a03456c80a19c14b9dc792c6f65df4a1d82d24e18b2c8bf4a4e6a2c86d
+size 382598708