Upload folder using huggingface_hub

train_vae.py

@@ -29,9 +29,9 @@ from collections import deque
 # --------------------------- Параметры ---------------------------
 ds_path            = "/workspace/d23"
 project            = "vae"
-batch_size         = 2
+batch_size         = 1
 base_learning_rate = 2e-6
-min_learning_rate  = 7e-7
+min_learning_rate  = 8e-8
 num_epochs         = 25
 sample_interval_share = 2
 use_wandb          = True
@@ -40,8 +40,8 @@ use_decay          = True
 optimizer_type     = "adam8bit"
 dtype              = torch.float32
 
-model_resolution   = 384
-high_resolution    = 768 
+model_resolution   = 512
+high_resolution    = 1024 
 limit              = 0
 save_barrier       = 1.3
 warmup_percent     = 0.001
@@ -50,7 +50,7 @@ beta2              = 0.997
 eps                = 1e-8
 clip_grad_norm     = 1.0
 mixed_precision    = "no"
-gradient_accumulation_steps = 8
+gradient_accumulation_steps = 16
 generated_folder   = "samples"
 save_as            = "vae"
 num_workers        = 0
@@ -169,9 +169,10 @@ if full_training and not train_decoder_only:
 else:
     # учим только 0-й блок декодера + post_quant_conv
     if hasattr(core, "decoder"):
-        # --- только 0-й up_block ---
         if hasattr(core.decoder, "up_blocks") and len(core.decoder.up_blocks) > 0:
-            for name, p in core.decoder.up_blocks[0].named_parameters():
+            # --- только 0-й up_block ---
+            # for name, p in core.decoder.up_blocks[0].named_parameters():
+            for name, p in core.decoder.up_blocks.named_parameters():
                 p.requires_grad = True
                 unfrozen_param_names.append(f"decoder.up_blocks.0.{name}")
         else:

vae/diffusion_pytorch_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bfd1075af9396f30739e76d965e79b5ec3daeedcd55392974e3835a3245922dc
+oid sha256:c917b96507d94e2c4954494c0479344574155ccbb479acbe0ba7a8c6f05e4af3
 size 444559412