Training in progress, step 123750

README.md

@@ -44,7 +44,7 @@ More information needed
 
 # Resource Usage Comparison
 
-- VRAM Use: 15.6991 GB
+- VRAM Use: 15.7006 GB
 
 # Distillation (Teacher -> Student) Architecture Difference:
 
@@ -85,7 +85,7 @@ Trained on 521,413,804 tokens from the [wikimedia/wikipedia](https://huggingface
 # Training Objective
 
 ```
-DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl), attn_loss_component=LossComponent(label=attn, weight=5, loss_fn=raw_mse, layer_mapper=layer-2, norm=layernorm_teacher_only, projector=mlp))
+DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl), attn_loss_component=LossComponent(label=attn, weight=5, loss_fn=raw_mse, layer_mapper=layer-2, norm=layernorm_teacher_only, projector=orthogonal))
 ```
 
 # Hyperparameters
@@ -94,16 +94,16 @@ The following hyperparameters were used during training:
 <details>
 <summary>Expand</summary>
 
-- learning_rate: `0.0002`
+- learning_rate: `0.0001`
 - train_batch_size: `8`
 - eval_batch_size: `8`
 - seed: `42`
 - optimizer: `Adam with betas=(0.9,0.999) and epsilon=1e-08`
 - lr_scheduler_type: `polynomial`
 - num_epochs: `1.0`
-- distillation_objective: `DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl), attn_loss_component=LossComponent(label=attn, weight=5, loss_fn=raw_mse, layer_mapper=layer-2, norm=layernorm_teacher_only, projector=mlp))`
+- distillation_objective: `DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl), attn_loss_component=LossComponent(label=attn, weight=5, loss_fn=raw_mse, layer_mapper=layer-2, norm=layernorm_teacher_only, projector=orthogonal))`
 - train_embeddings: `True`
-- lr_scheduler: `<torch.optim.lr_scheduler.LambdaLR object at 0x7f9ba0948f10>`
+- lr_scheduler: `<torch.optim.lr_scheduler.LambdaLR object at 0x7f3b60df7850>`
 - student_model_name_or_path: `None`
 - student_config_name_or_path: `distilbert/distilgpt2`
 - student_model_config: `None`
@@ -135,4 +135,4 @@ The following hyperparameters were used during training:
 - Distily 0.4.1
 - Transformers 4.44.2
 - Pytorch 2.4.0+cu121
-- Datasets 2.21.0
+- Datasets 2.18.0

logs/attn_norm=layernorm, attn_projector=orthogonal, attn_weight=5, learning_rate=0.0001, per_device_train_batch_size=8, warmup_ratio=0/events.out.tfevents.1725275867.a7e428977e35

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78cb22d18cf39cce6b81ce897e2dd5c0c69681e0acc15a5084e32cb4dbcb4624
+size 59393276

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c022ab2e5d29e956a4c5b5f6ae4b417b5a97cce6cd0ba8912338cf8bfc62a59e
+oid sha256:395c83ebb6e87c6467355754407035cbae9e55ebe7228cab9a04ebc10bea0cfc
 size 163832792

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7750bb1f3e7c5217dd7b95812e83f217a5b308efdadfdd2766fb79547cb33bcf
+oid sha256:8b84ade03e2b5b462154ae3fd3b1003abeeb5ef6c3abc2d996dc509e54f8e393
 size 5624