Training in progress, epoch 0

.ipynb_checkpoints/all_results-checkpoint.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 2.985781990521327,
+    "eval_accuracy": 0.23666666666666666,
+    "eval_loss": 2.182861566543579,
+    "eval_runtime": 72.7055,
+    "eval_samples_per_second": 82.525,
+    "eval_steps_per_second": 2.586,
+    "total_flos": 4.411646023570175e+19,
+    "train_loss": 2.258255742845081,
+    "train_runtime": 9823.5604,
+    "train_samples_per_second": 16.491,
+    "train_steps_per_second": 0.017
+}

.ipynb_checkpoints/eval_results-checkpoint.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.985781990521327,
+    "eval_accuracy": 0.23666666666666666,
+    "eval_loss": 2.182861566543579,
+    "eval_runtime": 72.7055,
+    "eval_samples_per_second": 82.525,
+    "eval_steps_per_second": 2.586
+}

.ipynb_checkpoints/train_results-checkpoint.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.985781990521327,
+    "total_flos": 4.411646023570175e+19,
+    "train_loss": 2.258255742845081,
+    "train_runtime": 9823.5604,
+    "train_samples_per_second": 16.491,
+    "train_steps_per_second": 0.017
+}

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 2.985781990521327,
-    "eval_accuracy": 0.43766666666666665,
-    "eval_loss": 2.0961079597473145,
-    "eval_runtime": 24.6417,
-    "eval_samples_per_second": 243.49,
-    "eval_steps_per_second": 1.907,
+    "eval_accuracy": 0.23666666666666666,
+    "eval_loss": 2.182861566543579,
+    "eval_runtime": 72.7055,
+    "eval_samples_per_second": 82.525,
+    "eval_steps_per_second": 2.586,
     "total_flos": 4.411646023570175e+19,
-    "train_loss": 2.214817982628232,
-    "train_runtime": 948.9642,
-    "train_samples_per_second": 170.712,
-    "train_steps_per_second": 0.332
+    "train_loss": 2.258255742845081,
+    "train_runtime": 9823.5604,
+    "train_samples_per_second": 16.491,
+    "train_steps_per_second": 0.017
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.985781990521327,
-    "eval_accuracy": 0.43766666666666665,
-    "eval_loss": 2.0961079597473145,
-    "eval_runtime": 24.6417,
-    "eval_samples_per_second": 243.49,
-    "eval_steps_per_second": 1.907
+    "eval_accuracy": 0.23666666666666666,
+    "eval_loss": 2.182861566543579,
+    "eval_runtime": 72.7055,
+    "eval_samples_per_second": 82.525,
+    "eval_steps_per_second": 2.586
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a33cf281bd713df6589f0072db8517a0a874cd790d92ce52838e3a0c4303efbf
+oid sha256:f8065d8c46ab562c99119416ffc302bbafeb72fe0b84f54d3baf4ec79435cc46
 size 1212808056

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 2.985781990521327,
     "total_flos": 4.411646023570175e+19,
-    "train_loss": 2.214817982628232,
-    "train_runtime": 948.9642,
-    "train_samples_per_second": 170.712,
-    "train_steps_per_second": 0.332
+    "train_loss": 2.258255742845081,
+    "train_runtime": 9823.5604,
+    "train_samples_per_second": 16.491,
+    "train_steps_per_second": 0.017
 }

trainer_state.json

@@ -1,269 +1,164 @@
 {
-  "best_metric": 0.43766666666666665,
-  "best_model_checkpoint": "clip-vit-large-patch14-finetuned-clip-vit-large-patch14-mnist_linear_probe/checkpoint-315",
+  "best_metric": 0.23666666666666666,
+  "best_model_checkpoint": "clip-vit-large-patch14-finetuned-clip-vit-large-patch14-mnist_linear_probe/checkpoint-168",
   "epoch": 2.985781990521327,
   "eval_steps": 500,
-  "global_step": 315,
+  "global_step": 168,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.0947867298578199,
-      "grad_norm": 2.288234233856201,
-      "learning_rate": 1.5625e-05,
-      "loss": 2.4094,
+      "epoch": 0.17772511848341233,
+      "grad_norm": 2.373176097869873,
+      "learning_rate": 2.9411764705882354e-05,
+      "loss": 2.4028,
       "step": 10
     },
     {
-      "epoch": 0.1895734597156398,
-      "grad_norm": 2.1556286811828613,
-      "learning_rate": 3.125e-05,
-      "loss": 2.3858,
+      "epoch": 0.35545023696682465,
+      "grad_norm": 2.2507705688476562,
+      "learning_rate": 4.900662251655629e-05,
+      "loss": 2.3848,
       "step": 20
     },
     {
-      "epoch": 0.2843601895734597,
-      "grad_norm": 2.1012516021728516,
-      "learning_rate": 4.6875e-05,
-      "loss": 2.3743,
+      "epoch": 0.533175355450237,
+      "grad_norm": 1.5781500339508057,
+      "learning_rate": 4.56953642384106e-05,
+      "loss": 2.3427,
       "step": 30
     },
     {
-      "epoch": 0.3791469194312796,
-      "grad_norm": 1.212685227394104,
-      "learning_rate": 4.858657243816255e-05,
-      "loss": 2.3353,
+      "epoch": 0.7109004739336493,
+      "grad_norm": 1.5205079317092896,
+      "learning_rate": 4.23841059602649e-05,
+      "loss": 2.3118,
       "step": 40
     },
     {
-      "epoch": 0.47393364928909953,
-      "grad_norm": 1.0641510486602783,
-      "learning_rate": 4.6819787985865726e-05,
-      "loss": 2.3048,
+      "epoch": 0.8886255924170616,
+      "grad_norm": 1.119712233543396,
+      "learning_rate": 3.907284768211921e-05,
+      "loss": 2.2886,
       "step": 50
     },
     {
-      "epoch": 0.5687203791469194,
-      "grad_norm": 0.9478774070739746,
-      "learning_rate": 4.50530035335689e-05,
-      "loss": 2.2905,
+      "epoch": 0.995260663507109,
+      "eval_accuracy": 0.157,
+      "eval_loss": 2.2661008834838867,
+      "eval_runtime": 478.4456,
+      "eval_samples_per_second": 12.541,
+      "eval_steps_per_second": 0.393,
+      "step": 56
+    },
+    {
+      "epoch": 1.066350710900474,
+      "grad_norm": 1.0753259658813477,
+      "learning_rate": 3.576158940397351e-05,
+      "loss": 2.2701,
       "step": 60
     },
     {
-      "epoch": 0.6635071090047393,
-      "grad_norm": 0.7033135294914246,
-      "learning_rate": 4.328621908127209e-05,
-      "loss": 2.2732,
+      "epoch": 1.2440758293838863,
+      "grad_norm": 0.8510423302650452,
+      "learning_rate": 3.2450331125827816e-05,
+      "loss": 2.2516,
       "step": 70
     },
     {
-      "epoch": 0.7582938388625592,
-      "grad_norm": 0.8783323168754578,
-      "learning_rate": 4.1519434628975266e-05,
-      "loss": 2.2636,
+      "epoch": 1.4218009478672986,
+      "grad_norm": 0.9210019111633301,
+      "learning_rate": 2.913907284768212e-05,
+      "loss": 2.2406,
       "step": 80
     },
     {
-      "epoch": 0.8530805687203792,
-      "grad_norm": 0.9178075790405273,
-      "learning_rate": 3.975265017667845e-05,
-      "loss": 2.2455,
+      "epoch": 1.599526066350711,
+      "grad_norm": 0.9463822245597839,
+      "learning_rate": 2.5827814569536424e-05,
+      "loss": 2.2312,
       "step": 90
     },
     {
-      "epoch": 0.9478672985781991,
-      "grad_norm": 0.6878641843795776,
-      "learning_rate": 3.7985865724381624e-05,
-      "loss": 2.2376,
+      "epoch": 1.7772511848341233,
+      "grad_norm": 0.8657775521278381,
+      "learning_rate": 2.2516556291390732e-05,
+      "loss": 2.2231,
       "step": 100
     },
     {
-      "epoch": 0.995260663507109,
-      "eval_accuracy": 0.2981666666666667,
-      "eval_loss": 2.2017133235931396,
-      "eval_runtime": 24.7638,
-      "eval_samples_per_second": 242.289,
-      "eval_steps_per_second": 1.898,
-      "step": 105
-    },
-    {
-      "epoch": 1.042654028436019,
-      "grad_norm": 0.9457613229751587,
-      "learning_rate": 3.621908127208481e-05,
-      "loss": 2.2256,
+      "epoch": 1.9549763033175356,
+      "grad_norm": 0.8078551292419434,
+      "learning_rate": 1.9205298013245036e-05,
+      "loss": 2.2153,
       "step": 110
     },
     {
-      "epoch": 1.1374407582938388,
-      "grad_norm": 0.850648820400238,
-      "learning_rate": 3.445229681978799e-05,
-      "loss": 2.2144,
+      "epoch": 1.9905213270142181,
+      "eval_accuracy": 0.1945,
+      "eval_loss": 2.2003655433654785,
+      "eval_runtime": 527.8354,
+      "eval_samples_per_second": 11.367,
+      "eval_steps_per_second": 0.356,
+      "step": 112
+    },
+    {
+      "epoch": 2.132701421800948,
+      "grad_norm": 0.7690733075141907,
+      "learning_rate": 1.589403973509934e-05,
+      "loss": 2.2121,
       "step": 120
     },
     {
-      "epoch": 1.2322274881516588,
-      "grad_norm": 0.8362263441085815,
-      "learning_rate": 3.2685512367491163e-05,
-      "loss": 2.2111,
+      "epoch": 2.31042654028436,
+      "grad_norm": 0.6478149890899658,
+      "learning_rate": 1.2582781456953644e-05,
+      "loss": 2.2047,
       "step": 130
     },
     {
-      "epoch": 1.3270142180094786,
-      "grad_norm": 0.6047300100326538,
-      "learning_rate": 3.0918727915194346e-05,
-      "loss": 2.2002,
+      "epoch": 2.4881516587677726,
+      "grad_norm": 0.6131073832511902,
+      "learning_rate": 9.271523178807948e-06,
+      "loss": 2.2018,
       "step": 140
     },
     {
-      "epoch": 1.4218009478672986,
-      "grad_norm": 0.9275925755500793,
-      "learning_rate": 2.915194346289753e-05,
-      "loss": 2.1941,
+      "epoch": 2.665876777251185,
+      "grad_norm": 0.7310860753059387,
+      "learning_rate": 5.9602649006622515e-06,
+      "loss": 2.2025,
       "step": 150
     },
-    {
-      "epoch": 1.5165876777251186,
-      "grad_norm": 0.8383633494377136,
-      "learning_rate": 2.738515901060071e-05,
-      "loss": 2.1916,
-      "step": 160
-    },
-    {
-      "epoch": 1.6113744075829384,
-      "grad_norm": 0.7653341293334961,
-      "learning_rate": 2.5618374558303885e-05,
-      "loss": 2.1859,
-      "step": 170
-    },
-    {
-      "epoch": 1.7061611374407581,
-      "grad_norm": 0.775774359703064,
-      "learning_rate": 2.3851590106007068e-05,
-      "loss": 2.1796,
-      "step": 180
-    },
-    {
-      "epoch": 1.8009478672985781,
-      "grad_norm": 0.932090699672699,
-      "learning_rate": 2.2084805653710246e-05,
-      "loss": 2.1708,
-      "step": 190
-    },
-    {
-      "epoch": 1.8957345971563981,
-      "grad_norm": 0.9772677421569824,
-      "learning_rate": 2.031802120141343e-05,
-      "loss": 2.1684,
-      "step": 200
-    },
-    {
-      "epoch": 1.9905213270142181,
-      "grad_norm": 0.7905783653259277,
-      "learning_rate": 1.855123674911661e-05,
-      "loss": 2.1641,
-      "step": 210
-    },
-    {
-      "epoch": 2.0,
-      "eval_accuracy": 0.4041666666666667,
-      "eval_loss": 2.1207876205444336,
-      "eval_runtime": 24.705,
-      "eval_samples_per_second": 242.866,
-      "eval_steps_per_second": 1.902,
-      "step": 211
-    },
-    {
-      "epoch": 2.085308056872038,
-      "grad_norm": 1.0608142614364624,
-      "learning_rate": 1.678445229681979e-05,
-      "loss": 2.1633,
-      "step": 220
-    },
-    {
-      "epoch": 2.1800947867298577,
-      "grad_norm": 1.0211544036865234,
-      "learning_rate": 1.501766784452297e-05,
-      "loss": 2.1518,
-      "step": 230
-    },
-    {
-      "epoch": 2.2748815165876777,
-      "grad_norm": 0.9483183026313782,
-      "learning_rate": 1.3250883392226149e-05,
-      "loss": 2.1507,
-      "step": 240
-    },
-    {
-      "epoch": 2.3696682464454977,
-      "grad_norm": 0.6434796452522278,
-      "learning_rate": 1.148409893992933e-05,
-      "loss": 2.1477,
-      "step": 250
-    },
-    {
-      "epoch": 2.4644549763033177,
-      "grad_norm": 0.7338018417358398,
-      "learning_rate": 9.717314487632508e-06,
-      "loss": 2.1439,
-      "step": 260
-    },
-    {
-      "epoch": 2.5592417061611377,
-      "grad_norm": 0.8191038966178894,
-      "learning_rate": 7.950530035335689e-06,
-      "loss": 2.1466,
-      "step": 270
-    },
-    {
-      "epoch": 2.654028436018957,
-      "grad_norm": 0.8383843898773193,
-      "learning_rate": 6.18374558303887e-06,
-      "loss": 2.1435,
-      "step": 280
-    },
-    {
-      "epoch": 2.748815165876777,
-      "grad_norm": 0.7335329055786133,
-      "learning_rate": 4.41696113074205e-06,
-      "loss": 2.1399,
-      "step": 290
-    },
     {
       "epoch": 2.843601895734597,
-      "grad_norm": 0.9002769589424133,
-      "learning_rate": 2.65017667844523e-06,
-      "loss": 2.142,
-      "step": 300
-    },
-    {
-      "epoch": 2.938388625592417,
-      "grad_norm": 0.8580403327941895,
-      "learning_rate": 8.8339222614841e-07,
-      "loss": 2.1453,
-      "step": 310
+      "grad_norm": 0.6716310381889343,
+      "learning_rate": 2.6490066225165563e-06,
+      "loss": 2.1981,
+      "step": 160
     },
     {
       "epoch": 2.985781990521327,
-      "eval_accuracy": 0.43766666666666665,
-      "eval_loss": 2.0961079597473145,
-      "eval_runtime": 24.9225,
-      "eval_samples_per_second": 240.747,
-      "eval_steps_per_second": 1.886,
-      "step": 315
+      "eval_accuracy": 0.23666666666666666,
+      "eval_loss": 2.182861566543579,
+      "eval_runtime": 73.356,
+      "eval_samples_per_second": 81.793,
+      "eval_steps_per_second": 2.563,
+      "step": 168
     },
     {
       "epoch": 2.985781990521327,
-      "step": 315,
+      "step": 168,
       "total_flos": 4.411646023570175e+19,
-      "train_loss": 2.214817982628232,
-      "train_runtime": 948.9642,
-      "train_samples_per_second": 170.712,
-      "train_steps_per_second": 0.332
+      "train_loss": 2.258255742845081,
+      "train_runtime": 9823.5604,
+      "train_samples_per_second": 16.491,
+      "train_steps_per_second": 0.017
     }
   ],
   "logging_steps": 10,
-  "max_steps": 315,
+  "max_steps": 168,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 500,
@@ -280,7 +175,7 @@
     }
   },
   "total_flos": 4.411646023570175e+19,
-  "train_batch_size": 128,
+  "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:281039b64ae59b55c230306a566c0dc20a32c69abd89b52f9831b2ba17247217
+oid sha256:ee5db6ddb9e63cd782c8fcac21a98b1f86c7908deeafe6f83803f0eab7db084c
 size 5304