lapp0 commited on
Commit
2f8a51b
·
verified ·
1 Parent(s): 75d6512

End of training

Browse files
README.md CHANGED
@@ -78,7 +78,7 @@ GPT2LMHeadModel(
78
 
79
  # Resource Usage
80
 
81
- - Max Train VRAM Use: 15.8655 GB
82
  - Available VRAM: 23.6429 GB
83
  - GPUs:
84
  - 1x NVIDIA GeForce RTX 4090
@@ -115,7 +115,7 @@ GPT2LMHeadModel(
115
  <br/>
116
 
117
  # Train Dataset
118
- Trained on 521,346,966 tokens from the [wikimedia/wikipedia](https://huggingface.co/datasets/wikimedia/wikipedia) dataset.
119
 
120
  - Num Samples: `990,000`
121
  - Subset: `20231101.en`
@@ -134,11 +134,7 @@ DistillationObjective(
134
  weight=0
135
  ),
136
  attn_loss_component=LossComponent(
137
- weight=5,
138
- loss_fn='raw_mse',
139
- layer_mapper='layer-2',
140
- norm='layernorm_teacher_only_affine',
141
- projector='mlp_64_l3'
142
  )
143
  )
144
  ```
@@ -165,14 +161,10 @@ The following hyperparameters were used during training:
165
  weight=0
166
  ),
167
  attn_loss_component=LossComponent(
168
- weight=5,
169
- loss_fn='raw_mse',
170
- layer_mapper='layer-2',
171
- norm='layernorm_teacher_only_affine',
172
- projector='mlp_64_l3'
173
  )
174
  )`
175
- - lr_scheduler: `<torch.optim.lr_scheduler.LambdaLR object at 0x7efc6c1dab60>`
176
  - student_model_name_or_path: `None`
177
  - student_config_name_or_path: `distilbert/distilgpt2`
178
  - student_model_config: `None`
@@ -191,7 +183,7 @@ The following hyperparameters were used during training:
191
  - gradient_accumulation_steps: `1`
192
  - weight_decay: `0.0`
193
  - max_grad_norm: `1.0`
194
- - warmup_ratio: `0`
195
  - warmup_steps: `0`
196
  - gradient_checkpointing: `True`
197
 
 
78
 
79
  # Resource Usage
80
 
81
+ - Max Train VRAM Use: 15.2844 GB
82
  - Available VRAM: 23.6429 GB
83
  - GPUs:
84
  - 1x NVIDIA GeForce RTX 4090
 
115
  <br/>
116
 
117
  # Train Dataset
118
+ Trained on 521,334,462 tokens from the [wikimedia/wikipedia](https://huggingface.co/datasets/wikimedia/wikipedia) dataset.
119
 
120
  - Num Samples: `990,000`
121
  - Subset: `20231101.en`
 
134
  weight=0
135
  ),
136
  attn_loss_component=LossComponent(
137
+ weight=0
 
 
 
 
138
  )
139
  )
140
  ```
 
161
  weight=0
162
  ),
163
  attn_loss_component=LossComponent(
164
+ weight=0
 
 
 
 
165
  )
166
  )`
167
+ - lr_scheduler: `<torch.optim.lr_scheduler.LambdaLR object at 0x7facb65ee170>`
168
  - student_model_name_or_path: `None`
169
  - student_config_name_or_path: `distilbert/distilgpt2`
170
  - student_model_config: `None`
 
183
  - gradient_accumulation_steps: `1`
184
  - weight_decay: `0.0`
185
  - max_grad_norm: `1.0`
186
+ - warmup_ratio: `0.0`
187
  - warmup_steps: `0`
188
  - gradient_checkpointing: `True`
189
 
logs/attn_weight=0/events.out.tfevents.1725646785.a7e428977e35 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:38cf51078356b96b7f32b5b97b17665b161655c3f6795053842fd60bd25928f9
3
- size 1910806
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:87b77783df13afe858bfb41cf0e02bb58a15d941fc29fb263a464be88df196e1
3
+ size 1970995
logs/attn_weight=0/events.out.tfevents.1725679647.a7e428977e35 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:79190e6f08b94a33384ec604313e0d71556e27915728ed6df60d6072c9633990
3
+ size 529
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:227716479b8200c41e9ec0b3caee412f709447e53ac129f1ff919dc0c1678c0d
3
  size 163832792
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c50903859f256cc52993c2d269c80e12457fead0c3de901f87ec1ee2082ca68f
3
  size 163832792