Training in progress, step 99000

README.md

@@ -44,7 +44,7 @@ More information needed
 
 # Resource Usage Comparison
 
-- VRAM Use: 7.4154 GB
+- VRAM Use: 7.4152 GB
 
 # Distillation (Teacher -> Student) Architecture Difference:
 
@@ -75,7 +75,7 @@ More information needed
 <br/>
 
 # Train Dataset
-Trained on 226,136,920 tokens from the [wikimedia/wikipedia](https://huggingface.co/datasets/wikimedia/wikipedia) dataset.
+Trained on 226,118,907 tokens from the [wikimedia/wikipedia](https://huggingface.co/datasets/wikimedia/wikipedia) dataset.
 
 - Num Samples: `396,000`
 - Subset: `20231101.en`
@@ -85,7 +85,7 @@ Trained on 226,136,920 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, 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, projector=orthogonal))
 ```
 
 # Hyperparameters
@@ -95,15 +95,16 @@ The following hyperparameters were used during training:
 <summary>Expand</summary>
 
 - learning_rate: `0.0002`
-- train_batch_size: `2`
+- train_batch_size: `4`
 - eval_batch_size: `8`
 - seed: `42`
 - optimizer: `Adam with betas=(0.9,0.999) and epsilon=1e-08`
 - lr_scheduler_type: `polynomial`
+- lr_scheduler_warmup_ratio: `0.2`
 - 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, 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, projector=orthogonal))`
 - train_embeddings: `True`
-- lr_scheduler: `<torch.optim.lr_scheduler.LambdaLR object at 0x7f6fd04f91b0>`
+- lr_scheduler: `<torch.optim.lr_scheduler.LambdaLR object at 0x7fe828138b50>`
 - student_model_name_or_path: `None`
 - student_config_name_or_path: `distilbert/distilgpt2`
 - student_model_config: `None`
@@ -123,7 +124,7 @@ The following hyperparameters were used during training:
 - gradient_accumulation_steps: `1`
 - weight_decay: `0.0`
 - max_grad_norm: `1.0`
-- warmup_ratio: `0`
+- warmup_ratio: `0.2`
 - warmup_steps: `0`
 - gradient_checkpointing: `True`
 
@@ -134,5 +135,5 @@ The following hyperparameters were used during training:
 # Framework Versions
 - Distily 0.4.1
 - Transformers 4.44.2
-- Pytorch 2.3.0
+- Pytorch 2.4.0+cu121
 - Datasets 2.21.0

logs/attn_norm=None, attn_projector=orthogonal, attn_weight=25, learning_rate=0.0001, per_device_train_batch_size=4, warmup_ratio=0.2/events.out.tfevents.1725101285.23668649e3db

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:909ef40f36715bdcf698d788f53666097f51d1ed0a550fc5da75b048aed42804
+size 47486052

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cfc887e1e01e9df85fea2c8bb05913c259ad49b5a551ed41d72ba3c4795f9671
+oid sha256:9ef7a40476007d3cc61e01ed4e06df0c4b2e8fe72ed3081f3e5cb306cab5cd71
 size 163832792

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9abad53848f3d5d36477a5bfac0a2ea2817c6265caa39e54846af2de0de58cc0
+oid sha256:07e795451ea41a525ff199bf7c3b971cb21ed8d822124b63f1fc03c6806654e4
 size 5560