Upload folder using huggingface_hub
Browse files- checkpoints/checkpoint-100.pt +2 -2
- checkpoints/checkpoint-25.pt +2 -2
- checkpoints/checkpoint-50.pt +2 -2
- checkpoints/checkpoint-75.pt +2 -2
- convert_checkpoints.py +1 -1
- model.yaml +1 -1
- training.yaml +1 -1
checkpoints/checkpoint-100.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d62df4a1682a61af8e5659f88f2acacc85a0893e05934765f8752c712ecd6dcc
|
| 3 |
+
size 2206381
|
checkpoints/checkpoint-25.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2bd4a8129d5b02bb12d9d4425cd79c5c9bccf4e69fcac4319cc8566d781d9a5a
|
| 3 |
+
size 2206345
|
checkpoints/checkpoint-50.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dc081f04306193f4ac19d96e50dd59ddfe8a2622c7a1875a436e417e6ffd9e57
|
| 3 |
+
size 2206345
|
checkpoints/checkpoint-75.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e6f5afd75cd02fdaa52989152fabc2123698d455925b98ab7b15713e518aa5ec
|
| 3 |
+
size 2206345
|
convert_checkpoints.py
CHANGED
|
@@ -30,7 +30,7 @@ for file in os.listdir(checkpoints_dir):
|
|
| 30 |
converted_state_dict[new_key] = value
|
| 31 |
|
| 32 |
# Save the converted checkpoint as a flat dictionary
|
| 33 |
-
output_path = os.path.join(checkpoints_dir, f"
|
| 34 |
torch.save(converted_state_dict, output_path)
|
| 35 |
print(f"Saved converted checkpoint to {output_path}")
|
| 36 |
else:
|
|
|
|
| 30 |
converted_state_dict[new_key] = value
|
| 31 |
|
| 32 |
# Save the converted checkpoint as a flat dictionary
|
| 33 |
+
output_path = os.path.join(checkpoints_dir, f"{file}")
|
| 34 |
torch.save(converted_state_dict, output_path)
|
| 35 |
print(f"Saved converted checkpoint to {output_path}")
|
| 36 |
else:
|
model.yaml
CHANGED
|
@@ -3,7 +3,7 @@ implementation: transformer_lens
|
|
| 3 |
model_name: default
|
| 4 |
n_layers: '2'
|
| 5 |
model_seed: '1'
|
| 6 |
-
d_model: '
|
| 7 |
n_ctx: '1024'
|
| 8 |
d_head: '2'
|
| 9 |
n_heads: '8'
|
|
|
|
| 3 |
model_name: default
|
| 4 |
n_layers: '2'
|
| 5 |
model_seed: '1'
|
| 6 |
+
d_model: '4'
|
| 7 |
n_ctx: '1024'
|
| 8 |
d_head: '2'
|
| 9 |
n_heads: '8'
|
training.yaml
CHANGED
|
@@ -28,7 +28,7 @@ warmup_steps: '0'
|
|
| 28 |
log_level: warning
|
| 29 |
log_level_replica: warning
|
| 30 |
log_on_each_node: 'True'
|
| 31 |
-
logging_dir: checkpoints/triangle-100k-og/runs/Jul09_16-
|
| 32 |
logging_strategy: IntervalStrategy.STEPS
|
| 33 |
logging_first_step: 'True'
|
| 34 |
logging_steps: '250'
|
|
|
|
| 28 |
log_level: warning
|
| 29 |
log_level_replica: warning
|
| 30 |
log_on_each_node: 'True'
|
| 31 |
+
logging_dir: checkpoints/triangle-100k-og/runs/Jul09_16-55-42_7be3271c880a
|
| 32 |
logging_strategy: IntervalStrategy.STEPS
|
| 33 |
logging_first_step: 'True'
|
| 34 |
logging_steps: '250'
|