Update params.yaml:train.epochs

dvc.lock

@@ -79,28 +79,28 @@ stages:
           output_folder: outs/train
           metrics_folder: outs/train_metrics
           pretrained_model: MoritzLaurer/xtremedistil-l6-h256-mnli-fever-anli-ling-binary
-          epochs: 8
+          epochs: 12
     outs:
     - path: outs/train
-      md5: 2d6f31d634d42ab9996d9cac216aff06.dir
-      size: 51992747
+      md5: 33778450a89f7e78533367bddb1b8a79.dir
+      size: 51985943
       nfiles: 7
     - path: outs/train_metrics
-      md5: 82d8979c72e4a138266128b85fdea5de.dir
-      size: 1652
+      md5: d917848924ece81fc408749b010704bf.dir
+      size: 2305
       nfiles: 10
     - path: outs/train_metrics.json
-      md5: 29e927a8537684c88395bd3a6fee69fe
-      size: 347
+      md5: c81edc01d691338e1561cd017c149691
+      size: 379
   eval:
     cmd: python src/eval.py outs/split/val.json outs/train outs/eval
     deps:
     - path: outs/split/val.json
       md5: 9f3681b095600e87d0bc922ecc902cc3
       size: 1216
     - path: outs/train
-      md5: 2d6f31d634d42ab9996d9cac216aff06.dir
-      size: 51992747
+      md5: 33778450a89f7e78533367bddb1b8a79.dir
+      size: 51985943
       nfiles: 7
     - path: src/eval.py
       md5: f733df5cfaaef682afcc65babff16c06
@@ -116,5 +116,5 @@ stages:
         - plots
     outs:
     - path: outs/eval/plots/confusion_matrix.json
-      md5: 7ad7650d9b00d4ae671f04ba20889b79
-      size: 1274
+      md5: d6b8bfebdf91a43b37079b9c87796a16
+      size: 1264

outs/eval/plots/confusion_matrix.json

@@ -5,23 +5,23 @@
     },
     {
         "actual": "data-sync",
-        "predicted": "experiments"
+        "predicted": "data-sync"
     },
     {
         "actual": "experiments",
         "predicted": "experiments"
     },
     {
         "actual": "data-sync",
-        "predicted": "experiments"
+        "predicted": "data-sync"
     },
     {
         "actual": "experiments",
         "predicted": "experiments"
     },
     {
         "actual": "data-sync",
-        "predicted": "experiments"
+        "predicted": "data-sync"
     },
     {
         "actual": "experiments",
@@ -49,7 +49,7 @@
     },
     {
         "actual": "data-sync",
-        "predicted": "experiments"
+        "predicted": "data-sync"
     },
     {
         "actual": "plots",
@@ -61,6 +61,6 @@
     },
     {
         "actual": "data-sync",
-        "predicted": "experiments"
+        "predicted": "data-sync"
     }
 ]

outs/train_metrics.json

@@ -1,13 +1,17 @@
 {
-    "step": 8,
-    "eval_loss": 0.9230659008026123,
-    "eval_accuracy": 0.5,
-    "eval_runtime": 0.0691,
-    "eval_samples_per_second": 231.714,
-    "eval_steps_per_second": 14.482,
-    "epoch": 8.0,
-    "train_runtime": 30.353,
-    "train_samples_per_second": 36.108,
-    "train_steps_per_second": 4.744,
-    "train_loss": 0.9423334333631728
+    "step": 12,
+    "eval": {
+        "loss": 0.5616878271102905,
+        "accuracy": 0.8125,
+        "runtime": 0.037,
+        "samples_per_second": 432.358,
+        "steps_per_second": 27.022
+    },
+    "epoch": 12.0,
+    "train": {
+        "runtime": 31.104,
+        "samples_per_second": 52.855,
+        "steps_per_second": 6.944,
+        "loss": 0.5529807761863426
+    }
 }

outs/train_metrics/scalars/epoch.tsv

@@ -1,10 +1,14 @@
 timestamp	step	epoch
-1656438398105	0	1.0
-1656438401893	1	2.0
-1656438405701	2	3.0
-1656438409437	3	4.0
-1656438413157	4	5.0
-1656438416948	5	6.0
-1656438420722	6	7.0
-1656438424590	7	8.0
-1656438424593	8	8.0
+1656499551214	0	1.0
+1656499553808	1	2.0
+1656499556420	2	3.0
+1656499559024	3	4.0
+1656499561643	4	5.0
+1656499564232	5	6.0
+1656499566822	6	7.0
+1656499569452	7	8.0
+1656499571980	8	9.0
+1656499574570	9	10.0
+1656499577134	10	11.0
+1656499579694	11	12.0
+1656499579697	12	12.0

outs/train_metrics/scalars/eval/accuracy.tsv

@@ -0,0 +1,13 @@
+timestamp	step	eval/accuracy
+1656499551214	0	0.5
+1656499553807	1	0.5
+1656499556418	2	0.5625
+1656499559023	3	0.625
+1656499561640	4	0.625
+1656499564229	5	0.8125
+1656499566821	6	0.8125
+1656499569451	7	0.8125
+1656499571977	8	0.8125
+1656499574563	9	0.8125
+1656499577133	10	0.8125
+1656499579692	11	0.8125

outs/train_metrics/scalars/eval/loss.tsv

@@ -0,0 +1,13 @@
+timestamp	step	eval/loss
+1656499551214	0	1.0209863185882568
+1656499553806	1	0.9715718626976013
+1656499556418	2	0.9561163783073425
+1656499559022	3	0.8461379408836365
+1656499561640	4	0.8161615133285522
+1656499564229	5	0.6735612750053406
+1656499566820	6	0.6104523539543152
+1656499569450	7	0.5978678464889526
+1656499571977	8	0.5740670561790466
+1656499574563	9	0.5696548819541931
+1656499577132	10	0.5634579062461853
+1656499579692	11	0.5616878271102905

outs/train_metrics/scalars/eval/runtime.tsv

@@ -0,0 +1,13 @@
+timestamp	step	eval/runtime
+1656499551214	0	0.0377
+1656499553807	1	0.037
+1656499556419	2	0.0368
+1656499559023	3	0.0375
+1656499561642	4	0.0375
+1656499564230	5	0.0367
+1656499566821	6	0.038
+1656499569451	7	0.0363
+1656499571978	8	0.039
+1656499574569	9	0.0384
+1656499577133	10	0.0362
+1656499579693	11	0.037

outs/train_metrics/scalars/eval/samples_per_second.tsv

@@ -0,0 +1,13 @@
+timestamp	step	eval/samples_per_second
+1656499551214	0	423.865
+1656499553808	1	432.868
+1656499556419	2	435.099
+1656499559023	3	427.211
+1656499561642	4	426.139
+1656499564230	5	435.721
+1656499566821	6	421.057
+1656499569452	7	440.376
+1656499571979	8	409.76
+1656499574570	9	416.347
+1656499577133	10	442.425
+1656499579693	11	432.358

outs/train_metrics/scalars/eval/steps_per_second.tsv

@@ -0,0 +1,13 @@
+timestamp	step	eval/steps_per_second
+1656499551214	0	26.492
+1656499553808	1	27.054
+1656499556419	2	27.194
+1656499559024	3	26.701
+1656499561642	4	26.634
+1656499564231	5	27.233
+1656499566822	6	26.316
+1656499569452	7	27.523
+1656499571979	8	25.61
+1656499574570	9	26.022
+1656499577134	10	27.652
+1656499579694	11	27.022

outs/train_metrics/scalars/train/loss.tsv

@@ -0,0 +1,2 @@
+timestamp	step	train/loss
+1656499579697	12	0.5529807761863426

outs/train_metrics/scalars/train/runtime.tsv

@@ -0,0 +1,2 @@
+timestamp	step	train/runtime
+1656499579695	12	31.104

outs/train_metrics/scalars/train/samples_per_second.tsv

@@ -0,0 +1,2 @@
+timestamp	step	train/samples_per_second
+1656499579695	12	52.855

outs/train_metrics/scalars/train/steps_per_second.tsv

@@ -0,0 +1,2 @@
+timestamp	step	train/steps_per_second
+1656499579696	12	6.944

params.yaml

@@ -21,7 +21,7 @@ train:
   output_folder: outs/train
   metrics_folder: outs/train_metrics
   pretrained_model: MoritzLaurer/xtremedistil-l6-h256-mnli-fever-anli-ling-binary
-  epochs: 8
+  epochs: 12
 
 eval:
   output_folder: outs/eval