Training in progress, step 16000
Browse files- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/pytorch_model.bin +1 -1
- last-checkpoint/rng_state.pth +1 -1
- last-checkpoint/scaler.pt +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +21 -3
- pytorch_model.bin +1 -1
last-checkpoint/optimizer.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 484079855
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e7362a3aee96b4023fb719b627d28c7af2ad7d5bd85777d2ac201d1a03671c58
|
3 |
size 484079855
|
last-checkpoint/pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 242029307
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7676f22cc8f3e82775b885956c2ec97602d93e20c376d5e6a7cd17517102e0c4
|
3 |
size 242029307
|
last-checkpoint/rng_state.pth
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 14503
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bbcf03047cce20a73278edccfd55c1e16299a79e06a1b14b5dbf7b7dfb86be69
|
3 |
size 14503
|
last-checkpoint/scaler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 559
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4d8decf4db2600bb739bf3dce5a704a984d64dfc520d417c2b991d3b763cf6bb
|
3 |
size 559
|
last-checkpoint/scheduler.pt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 623
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1007a894fa8cc877a251e1ca9cf115dad8fb79437a615b8df18308ab041a0b78
|
3 |
size 623
|
last-checkpoint/trainer_state.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"best_metric": 1.5166430473327637,
|
3 |
"best_model_checkpoint": "/home2/s5431786/jb-internship-repr/results/Salesforce-codet5-small-java-small-selected/checkpoint-2000",
|
4 |
-
"epoch":
|
5 |
-
"global_step":
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
@@ -132,11 +132,29 @@
|
|
132 |
"eval_samples_per_second": 143.768,
|
133 |
"eval_steps_per_second": 1.501,
|
134 |
"step": 14000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
135 |
}
|
136 |
],
|
137 |
"max_steps": 72090,
|
138 |
"num_train_epochs": 10,
|
139 |
-
"total_flos":
|
140 |
"trial_name": null,
|
141 |
"trial_params": null
|
142 |
}
|
|
|
1 |
{
|
2 |
"best_metric": 1.5166430473327637,
|
3 |
"best_model_checkpoint": "/home2/s5431786/jb-internship-repr/results/Salesforce-codet5-small-java-small-selected/checkpoint-2000",
|
4 |
+
"epoch": 2.2194479123318076,
|
5 |
+
"global_step": 16000,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
|
|
132 |
"eval_samples_per_second": 143.768,
|
133 |
"eval_steps_per_second": 1.501,
|
134 |
"step": 14000
|
135 |
+
},
|
136 |
+
{
|
137 |
+
"epoch": 2.22,
|
138 |
+
"learning_rate": 8.191136745272688e-05,
|
139 |
+
"loss": 0.9368,
|
140 |
+
"step": 16000
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 2.22,
|
144 |
+
"eval_exact_match": 0.318906223787955,
|
145 |
+
"eval_loss": 1.6294344663619995,
|
146 |
+
"eval_rouge1": 0.5072474931474986,
|
147 |
+
"eval_rouge2": 0.2326535102579398,
|
148 |
+
"eval_rougeL": 0.5071239367268146,
|
149 |
+
"eval_runtime": 165.6048,
|
150 |
+
"eval_samples_per_second": 143.981,
|
151 |
+
"eval_steps_per_second": 1.504,
|
152 |
+
"step": 16000
|
153 |
}
|
154 |
],
|
155 |
"max_steps": 72090,
|
156 |
"num_train_epochs": 10,
|
157 |
+
"total_flos": 2.0786064553672704e+17,
|
158 |
"trial_name": null,
|
159 |
"trial_params": null
|
160 |
}
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 242029307
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7676f22cc8f3e82775b885956c2ec97602d93e20c376d5e6a7cd17517102e0c4
|
3 |
size 242029307
|