Training in progress, step 10000, checkpoint

checkpoint-10000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:293d2a96d11d48530f2f9b8f44f37eb2ddd94375961f9d9e6a64e49d96a5a41e
+oid sha256:075aa760d6fd5f4c80ba5e73271f675bc98eea9b346a3e3ef66739e9358d001f
 size 1711848436

checkpoint-10000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:35d7a235dedb65277f5c4e656fae5e49ec7ea33c0cc7898c52db72b626b08486
+oid sha256:58a4bc9ece8d959bb61198b772f95d1475943d58521c42e37284cf83b5e461a9
 size 3424043887

checkpoint-10000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e61e6b2c18ca2fcc4852ae87d1212fb04c9680cdf4abad489ed58d5aaf87cd4a
+oid sha256:fd1f9483f6c91649db661b676a4233e1f965a0ca5eb25be5376d69cc5dfe4ae0
 size 623

checkpoint-10000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.9420022368431091,
+  "best_metric": 0.9393072724342346,
   "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l58-l/checkpoint-10000",
-  "epoch": 1.841959845275373,
+  "epoch": 1.8422991893883567,
   "eval_steps": 500,
   "global_step": 10000,
   "is_hyper_param_search": false,
@@ -9,308 +9,308 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.09209799226376865,
-      "grad_norm": 0.0014697719598188996,
-      "learning_rate": 4.884877509670289e-08,
-      "loss": 0.4588,
+      "epoch": 0.09211495946941783,
+      "grad_norm": 20.04043960571289,
+      "learning_rate": 4.8848563006632274e-08,
+      "loss": 0.3758,
       "step": 500
     },
     {
-      "epoch": 0.09209799226376865,
-      "eval_loss": 1.4194883108139038,
-      "eval_runtime": 78.7703,
-      "eval_samples_per_second": 15.31,
-      "eval_steps_per_second": 1.917,
+      "epoch": 0.09211495946941783,
+      "eval_loss": 1.4184927940368652,
+      "eval_runtime": 74.3562,
+      "eval_samples_per_second": 16.233,
+      "eval_steps_per_second": 2.031,
       "step": 500
     },
     {
-      "epoch": 0.1841959845275373,
-      "grad_norm": 301.76043701171875,
-      "learning_rate": 4.769755019340578e-08,
-      "loss": 0.4255,
+      "epoch": 0.18422991893883567,
+      "grad_norm": 74.25025939941406,
+      "learning_rate": 4.7697126013264556e-08,
+      "loss": 0.4103,
       "step": 1000
     },
     {
-      "epoch": 0.1841959845275373,
-      "eval_loss": 1.3416553735733032,
-      "eval_runtime": 77.0051,
-      "eval_samples_per_second": 15.661,
-      "eval_steps_per_second": 1.961,
+      "epoch": 0.18422991893883567,
+      "eval_loss": 1.3500770330429077,
+      "eval_runtime": 75.7013,
+      "eval_samples_per_second": 15.944,
+      "eval_steps_per_second": 1.995,
       "step": 1000
     },
     {
-      "epoch": 0.27629397679130596,
-      "grad_norm": 35.28853225708008,
-      "learning_rate": 4.654632529010867e-08,
-      "loss": 0.3724,
+      "epoch": 0.2763448784082535,
+      "grad_norm": 0.5102410912513733,
+      "learning_rate": 4.6545689019896826e-08,
+      "loss": 0.433,
       "step": 1500
     },
     {
-      "epoch": 0.27629397679130596,
-      "eval_loss": 1.2872973680496216,
-      "eval_runtime": 77.5439,
-      "eval_samples_per_second": 15.552,
-      "eval_steps_per_second": 1.947,
+      "epoch": 0.2763448784082535,
+      "eval_loss": 1.2885024547576904,
+      "eval_runtime": 75.197,
+      "eval_samples_per_second": 16.051,
+      "eval_steps_per_second": 2.008,
       "step": 1500
     },
     {
-      "epoch": 0.3683919690550746,
-      "grad_norm": 338.9349060058594,
-      "learning_rate": 4.539510038681156e-08,
-      "loss": 0.3251,
+      "epoch": 0.36845983787767134,
+      "grad_norm": 0.1667678952217102,
+      "learning_rate": 4.539425202652911e-08,
+      "loss": 0.3424,
       "step": 2000
     },
     {
-      "epoch": 0.3683919690550746,
-      "eval_loss": 1.234910249710083,
-      "eval_runtime": 78.7715,
-      "eval_samples_per_second": 15.31,
-      "eval_steps_per_second": 1.917,
+      "epoch": 0.36845983787767134,
+      "eval_loss": 1.239119052886963,
+      "eval_runtime": 75.387,
+      "eval_samples_per_second": 16.011,
+      "eval_steps_per_second": 2.003,
       "step": 2000
     },
     {
-      "epoch": 0.46048996131884323,
-      "grad_norm": 325.85101318359375,
-      "learning_rate": 4.4243875483514457e-08,
-      "loss": 0.3308,
+      "epoch": 0.46057479734708917,
+      "grad_norm": 343.8175048828125,
+      "learning_rate": 4.4242815033161385e-08,
+      "loss": 0.3645,
       "step": 2500
     },
     {
-      "epoch": 0.46048996131884323,
-      "eval_loss": 1.1944907903671265,
-      "eval_runtime": 79.0397,
-      "eval_samples_per_second": 15.258,
-      "eval_steps_per_second": 1.91,
+      "epoch": 0.46057479734708917,
+      "eval_loss": 1.1902339458465576,
+      "eval_runtime": 75.9899,
+      "eval_samples_per_second": 15.884,
+      "eval_steps_per_second": 1.987,
       "step": 2500
     },
     {
-      "epoch": 0.5525879535826119,
-      "grad_norm": 100.85855102539062,
-      "learning_rate": 4.309265058021735e-08,
-      "loss": 0.3017,
+      "epoch": 0.552689756816507,
+      "grad_norm": 305.5626525878906,
+      "learning_rate": 4.309137803979366e-08,
+      "loss": 0.3172,
       "step": 3000
     },
     {
-      "epoch": 0.5525879535826119,
-      "eval_loss": 1.1593009233474731,
-      "eval_runtime": 77.6354,
-      "eval_samples_per_second": 15.534,
-      "eval_steps_per_second": 1.945,
+      "epoch": 0.552689756816507,
+      "eval_loss": 1.1506118774414062,
+      "eval_runtime": 74.8655,
+      "eval_samples_per_second": 16.122,
+      "eval_steps_per_second": 2.017,
       "step": 3000
     },
     {
-      "epoch": 0.6446859458463805,
-      "grad_norm": 8.865598678588867,
-      "learning_rate": 4.194142567692024e-08,
-      "loss": 0.2962,
+      "epoch": 0.6448047162859248,
+      "grad_norm": 65.59048461914062,
+      "learning_rate": 4.193994104642594e-08,
+      "loss": 0.2751,
       "step": 3500
     },
     {
-      "epoch": 0.6446859458463805,
-      "eval_loss": 1.1259427070617676,
-      "eval_runtime": 78.2715,
-      "eval_samples_per_second": 15.408,
-      "eval_steps_per_second": 1.929,
+      "epoch": 0.6448047162859248,
+      "eval_loss": 1.1169357299804688,
+      "eval_runtime": 76.2936,
+      "eval_samples_per_second": 15.82,
+      "eval_steps_per_second": 1.979,
       "step": 3500
     },
     {
-      "epoch": 0.7367839381101492,
-      "grad_norm": 4.393447399139404,
-      "learning_rate": 4.079020077362314e-08,
+      "epoch": 0.7369196757553427,
+      "grad_norm": 18.106201171875,
+      "learning_rate": 4.078850405305821e-08,
       "loss": 0.2919,
       "step": 4000
     },
     {
-      "epoch": 0.7367839381101492,
-      "eval_loss": 1.0954023599624634,
-      "eval_runtime": 78.5366,
-      "eval_samples_per_second": 15.356,
-      "eval_steps_per_second": 1.923,
+      "epoch": 0.7369196757553427,
+      "eval_loss": 1.0920671224594116,
+      "eval_runtime": 74.9101,
+      "eval_samples_per_second": 16.113,
+      "eval_steps_per_second": 2.016,
       "step": 4000
     },
     {
-      "epoch": 0.8288819303739179,
-      "grad_norm": 1.6036031246185303,
-      "learning_rate": 3.9638975870326026e-08,
-      "loss": 0.307,
+      "epoch": 0.8290346352247605,
+      "grad_norm": 506.40570068359375,
+      "learning_rate": 3.9637067059690496e-08,
+      "loss": 0.2583,
       "step": 4500
     },
     {
-      "epoch": 0.8288819303739179,
-      "eval_loss": 1.0728861093521118,
-      "eval_runtime": 78.4032,
-      "eval_samples_per_second": 15.382,
-      "eval_steps_per_second": 1.926,
+      "epoch": 0.8290346352247605,
+      "eval_loss": 1.0721209049224854,
+      "eval_runtime": 75.8053,
+      "eval_samples_per_second": 15.922,
+      "eval_steps_per_second": 1.992,
       "step": 4500
     },
     {
-      "epoch": 0.9209799226376865,
-      "grad_norm": 11.990758895874023,
-      "learning_rate": 3.8487750967028915e-08,
-      "loss": 0.2764,
+      "epoch": 0.9211495946941783,
+      "grad_norm": 1.1508910655975342,
+      "learning_rate": 3.848563006632277e-08,
+      "loss": 0.2679,
       "step": 5000
     },
     {
-      "epoch": 0.9209799226376865,
-      "eval_loss": 1.0523593425750732,
-      "eval_runtime": 78.3726,
-      "eval_samples_per_second": 15.388,
-      "eval_steps_per_second": 1.927,
+      "epoch": 0.9211495946941783,
+      "eval_loss": 1.0519349575042725,
+      "eval_runtime": 75.226,
+      "eval_samples_per_second": 16.045,
+      "eval_steps_per_second": 2.007,
       "step": 5000
     },
     {
-      "epoch": 1.013077914901455,
-      "grad_norm": 43.865562438964844,
-      "learning_rate": 3.733652606373181e-08,
-      "loss": 0.2456,
+      "epoch": 1.013264554163596,
+      "grad_norm": 0.1393290013074875,
+      "learning_rate": 3.733419307295505e-08,
+      "loss": 0.2472,
       "step": 5500
     },
     {
-      "epoch": 1.013077914901455,
-      "eval_loss": 1.0375442504882812,
-      "eval_runtime": 78.2473,
-      "eval_samples_per_second": 15.413,
-      "eval_steps_per_second": 1.93,
+      "epoch": 1.013264554163596,
+      "eval_loss": 1.0355629920959473,
+      "eval_runtime": 74.6535,
+      "eval_samples_per_second": 16.168,
+      "eval_steps_per_second": 2.023,
       "step": 5500
     },
     {
-      "epoch": 1.1051759071652238,
-      "grad_norm": 105.91036987304688,
-      "learning_rate": 3.61853011604347e-08,
-      "loss": 0.2642,
+      "epoch": 1.105379513633014,
+      "grad_norm": 200.34664916992188,
+      "learning_rate": 3.6182756079587324e-08,
+      "loss": 0.26,
       "step": 6000
     },
     {
-      "epoch": 1.1051759071652238,
-      "eval_loss": 1.0233356952667236,
-      "eval_runtime": 78.3135,
-      "eval_samples_per_second": 15.4,
-      "eval_steps_per_second": 1.928,
+      "epoch": 1.105379513633014,
+      "eval_loss": 1.0176944732666016,
+      "eval_runtime": 74.4374,
+      "eval_samples_per_second": 16.215,
+      "eval_steps_per_second": 2.029,
       "step": 6000
     },
     {
-      "epoch": 1.1972738994289924,
-      "grad_norm": 221.4896240234375,
-      "learning_rate": 3.503407625713759e-08,
-      "loss": 0.2066,
+      "epoch": 1.1974944731024317,
+      "grad_norm": 431.1804504394531,
+      "learning_rate": 3.50313190862196e-08,
+      "loss": 0.2153,
       "step": 6500
     },
     {
-      "epoch": 1.1972738994289924,
-      "eval_loss": 1.0104238986968994,
-      "eval_runtime": 78.6485,
-      "eval_samples_per_second": 15.334,
-      "eval_steps_per_second": 1.92,
+      "epoch": 1.1974944731024317,
+      "eval_loss": 1.0044586658477783,
+      "eval_runtime": 74.3247,
+      "eval_samples_per_second": 16.24,
+      "eval_steps_per_second": 2.032,
       "step": 6500
     },
     {
-      "epoch": 1.289371891692761,
-      "grad_norm": 270.53253173828125,
-      "learning_rate": 3.3882851353840485e-08,
-      "loss": 0.2376,
+      "epoch": 1.2896094325718497,
+      "grad_norm": 0.6204941272735596,
+      "learning_rate": 3.387988209285188e-08,
+      "loss": 0.1791,
       "step": 7000
     },
     {
-      "epoch": 1.289371891692761,
-      "eval_loss": 0.998353898525238,
-      "eval_runtime": 78.5502,
-      "eval_samples_per_second": 15.353,
-      "eval_steps_per_second": 1.922,
+      "epoch": 1.2896094325718497,
+      "eval_loss": 0.9926707148551941,
+      "eval_runtime": 75.7317,
+      "eval_samples_per_second": 15.938,
+      "eval_steps_per_second": 1.994,
       "step": 7000
     },
     {
-      "epoch": 1.3814698839565298,
-      "grad_norm": 402.9339904785156,
-      "learning_rate": 3.2731626450543374e-08,
-      "loss": 0.1931,
+      "epoch": 1.3817243920412676,
+      "grad_norm": 52.27273178100586,
+      "learning_rate": 3.272844509948415e-08,
+      "loss": 0.2082,
       "step": 7500
     },
     {
-      "epoch": 1.3814698839565298,
-      "eval_loss": 0.9887453317642212,
-      "eval_runtime": 78.4833,
-      "eval_samples_per_second": 15.366,
-      "eval_steps_per_second": 1.924,
+      "epoch": 1.3817243920412676,
+      "eval_loss": 0.9803809523582458,
+      "eval_runtime": 74.9936,
+      "eval_samples_per_second": 16.095,
+      "eval_steps_per_second": 2.014,
       "step": 7500
     },
     {
-      "epoch": 1.4735678762202984,
-      "grad_norm": 0.006355441175401211,
-      "learning_rate": 3.158040154724626e-08,
-      "loss": 0.2163,
+      "epoch": 1.4738393515106853,
+      "grad_norm": 0.0976635068655014,
+      "learning_rate": 3.1577008106116435e-08,
+      "loss": 0.196,
       "step": 8000
     },
     {
-      "epoch": 1.4735678762202984,
-      "eval_loss": 0.9767189621925354,
-      "eval_runtime": 78.5499,
-      "eval_samples_per_second": 15.353,
-      "eval_steps_per_second": 1.922,
+      "epoch": 1.4738393515106853,
+      "eval_loss": 0.9711655378341675,
+      "eval_runtime": 74.3537,
+      "eval_samples_per_second": 16.233,
+      "eval_steps_per_second": 2.031,
       "step": 8000
     },
     {
-      "epoch": 1.565665868484067,
-      "grad_norm": 0.12982100248336792,
-      "learning_rate": 3.042917664394916e-08,
-      "loss": 0.1903,
+      "epoch": 1.565954310980103,
+      "grad_norm": 46.64162826538086,
+      "learning_rate": 3.042557111274871e-08,
+      "loss": 0.1946,
       "step": 8500
     },
     {
-      "epoch": 1.565665868484067,
-      "eval_loss": 0.9664921760559082,
-      "eval_runtime": 78.6832,
-      "eval_samples_per_second": 15.327,
-      "eval_steps_per_second": 1.919,
+      "epoch": 1.565954310980103,
+      "eval_loss": 0.9621181488037109,
+      "eval_runtime": 75.7205,
+      "eval_samples_per_second": 15.94,
+      "eval_steps_per_second": 1.994,
       "step": 8500
     },
     {
-      "epoch": 1.6577638607478358,
-      "grad_norm": 173.31788635253906,
-      "learning_rate": 2.927795174065205e-08,
-      "loss": 0.2069,
+      "epoch": 1.658069270449521,
+      "grad_norm": 0.05690698325634003,
+      "learning_rate": 2.927413411938099e-08,
+      "loss": 0.2422,
       "step": 9000
     },
     {
-      "epoch": 1.6577638607478358,
-      "eval_loss": 0.9571623206138611,
-      "eval_runtime": 78.544,
-      "eval_samples_per_second": 15.354,
-      "eval_steps_per_second": 1.922,
+      "epoch": 1.658069270449521,
+      "eval_loss": 0.9536784291267395,
+      "eval_runtime": 75.1745,
+      "eval_samples_per_second": 16.056,
+      "eval_steps_per_second": 2.009,
       "step": 9000
     },
     {
-      "epoch": 1.7498618530116044,
-      "grad_norm": 358.7403259277344,
-      "learning_rate": 2.8126726837354947e-08,
-      "loss": 0.2093,
+      "epoch": 1.750184229918939,
+      "grad_norm": 0.08470064401626587,
+      "learning_rate": 2.8122697126013263e-08,
+      "loss": 0.2106,
       "step": 9500
     },
     {
-      "epoch": 1.7498618530116044,
-      "eval_loss": 0.9496769905090332,
-      "eval_runtime": 78.3124,
-      "eval_samples_per_second": 15.4,
-      "eval_steps_per_second": 1.928,
+      "epoch": 1.750184229918939,
+      "eval_loss": 0.9457550644874573,
+      "eval_runtime": 74.5062,
+      "eval_samples_per_second": 16.2,
+      "eval_steps_per_second": 2.027,
       "step": 9500
     },
     {
-      "epoch": 1.841959845275373,
-      "grad_norm": 435.9449768066406,
-      "learning_rate": 2.6975501934057836e-08,
-      "loss": 0.2523,
+      "epoch": 1.8422991893883567,
+      "grad_norm": 1.5745967626571655,
+      "learning_rate": 2.697126013264554e-08,
+      "loss": 0.1801,
       "step": 10000
     },
     {
-      "epoch": 1.841959845275373,
-      "eval_loss": 0.9420022368431091,
-      "eval_runtime": 78.3702,
-      "eval_samples_per_second": 15.388,
-      "eval_steps_per_second": 1.927,
+      "epoch": 1.8422991893883567,
+      "eval_loss": 0.9393072724342346,
+      "eval_runtime": 75.9432,
+      "eval_samples_per_second": 15.893,
+      "eval_steps_per_second": 1.988,
       "step": 10000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 21716,
+  "max_steps": 21712,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 4,
   "save_steps": 500,
@@ -326,7 +326,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 3598123150126890.0,
+  "total_flos": 3598303029302520.0,
   "train_batch_size": 2,
   "trial_name": null,
   "trial_params": null

checkpoint-10000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:417b0bacc3a8958dfa75621764230fb890db48c09ca241dff24e5264d406285b
+oid sha256:9a29c2374e59ce3f57b34db9cb20d2d1bdc870c3dad128e9a4d462636c37e8b0
 size 4847