Upload RT-DETRv2 voucher classifier

README.md

@@ -86,8 +86,8 @@ This model is a fine-tuned version of [PekingU/rtdetr_v2_r101vd](https://hugging
 - **Tesoreria receipts**: 0.0000 (needs improvement)
 
 **Model Confidence:**
-- **Digital invoices mean confidence**: 0.5811 (low)
-- **Fisico receipts mean confidence**: 0.0000 (low)
+- **Digital invoices mean confidence**: 0.4218 (low)
+- **Fisico receipts mean confidence**: 0.3837 (low)
 - **Tesoreria receipts mean confidence**: 0.0000 (low)
 
 **Performance by Object Size:**
@@ -114,16 +114,16 @@ This model is a fine-tuned version of [PekingU/rtdetr_v2_r101vd](https://hugging
 - **RAM**: 235.9 GB
 - **GPU configuration**: H100 optimized
 
-**Training Time**: 27.6 minutes
+**Training Time**: 27.0 minutes
 
 **Training Summary:**
-- **Final training loss**: 10.5270
+- **Final training loss**: 10.7460
 - **Final learning rate**: 1.77e-11
 
 
 ### MLflow Tracking
 
-- **MLflow Run ID**: 27297b63155142db8c9b015a3205e499
+- **MLflow Run ID**: 6b50f63a6e3144b7a719bbb2b15cb77a
 - **MLflow Experiment**: RT-DETRv2_Voucher_Classification
 
 

checkpoint-1000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8e15219ede20064681b971044da73d7daf73eec755a5acb92dea3cf8158fc8d9
+oid sha256:e8016ef1056593c544ea3cc47df3ed4b46437434f1c167bd7c17e98362bf291e
 size 306699044

checkpoint-1000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0297a542a16615f183c62cb889e5a7f4d3b07fedbf590dac7021616a59cf1d03
+oid sha256:ecf9a17d55473fe89a13c7be95a701c1739b4ab392380bc62fe3414b18a61724
 size 611580898

checkpoint-1000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_global_step": 1000,
-  "best_metric": 65.57749938964844,
-  "best_model_checkpoint": "rtdetr_voucher_model/checkpoint-1000",
+  "best_global_step": 800,
+  "best_metric": 75.64146423339844,
+  "best_model_checkpoint": "rtdetr_voucher_model/checkpoint-800",
   "epoch": 47.61904761904762,
   "eval_steps": 200,
   "global_step": 1000,
@@ -11,217 +11,217 @@
   "log_history": [
     {
       "epoch": 1.9047619047619047,
-      "grad_norm": 10800.818359375,
+      "grad_norm": 23553.73828125,
       "learning_rate": 7.8e-07,
-      "loss": 1388.4377,
+      "loss": 1390.8067,
       "step": 40
     },
     {
       "epoch": 3.8095238095238093,
-      "grad_norm": 21733.138671875,
+      "grad_norm": 7427.158203125,
       "learning_rate": 1.5800000000000001e-06,
-      "loss": 1273.0859,
+      "loss": 1273.2542,
       "step": 80
     },
     {
       "epoch": 5.714285714285714,
-      "grad_norm": 2858.446044921875,
+      "grad_norm": 6636.22314453125,
       "learning_rate": 2.38e-06,
-      "loss": 1063.7887,
+      "loss": 1059.9522,
       "step": 120
     },
     {
       "epoch": 7.619047619047619,
-      "grad_norm": 4524.0263671875,
+      "grad_norm": 1760.4281005859375,
       "learning_rate": 3.1800000000000005e-06,
-      "loss": 771.9299,
+      "loss": 790.0425,
       "step": 160
     },
     {
       "epoch": 9.523809523809524,
-      "grad_norm": 2449.484130859375,
+      "grad_norm": 1625.7744140625,
       "learning_rate": 3.980000000000001e-06,
-      "loss": 519.636,
+      "loss": 517.1978,
       "step": 200
     },
     {
       "epoch": 9.523809523809524,
-      "eval_loss": 561.4364624023438,
-      "eval_runtime": 8.2677,
-      "eval_samples_per_second": 14.03,
-      "eval_steps_per_second": 0.484,
+      "eval_loss": 535.0709838867188,
+      "eval_runtime": 8.3595,
+      "eval_samples_per_second": 13.876,
+      "eval_steps_per_second": 0.478,
       "step": 200
     },
     {
       "epoch": 11.428571428571429,
-      "grad_norm": 3173.60546875,
+      "grad_norm": 1354.9310302734375,
       "learning_rate": 4.78e-06,
-      "loss": 396.2762,
+      "loss": 353.8703,
       "step": 240
     },
     {
       "epoch": 13.333333333333334,
-      "grad_norm": 957.3853149414062,
+      "grad_norm": 873.0985107421875,
       "learning_rate": 5.580000000000001e-06,
-      "loss": 317.4443,
+      "loss": 262.4233,
       "step": 280
     },
     {
       "epoch": 15.238095238095237,
-      "grad_norm": 2215.865966796875,
+      "grad_norm": 1516.567626953125,
       "learning_rate": 6.380000000000001e-06,
-      "loss": 244.5671,
+      "loss": 205.628,
       "step": 320
     },
     {
       "epoch": 17.142857142857142,
-      "grad_norm": 494.6834716796875,
+      "grad_norm": 5139.3779296875,
       "learning_rate": 7.180000000000001e-06,
-      "loss": 200.2045,
+      "loss": 174.0845,
       "step": 360
     },
     {
       "epoch": 19.047619047619047,
-      "grad_norm": 389.99395751953125,
+      "grad_norm": 480.6575622558594,
       "learning_rate": 7.980000000000002e-06,
-      "loss": 143.4672,
+      "loss": 142.4243,
       "step": 400
     },
     {
       "epoch": 19.047619047619047,
-      "eval_loss": 212.24693298339844,
-      "eval_runtime": 7.6261,
-      "eval_samples_per_second": 15.211,
-      "eval_steps_per_second": 0.525,
+      "eval_loss": 214.48715209960938,
+      "eval_runtime": 7.9717,
+      "eval_samples_per_second": 14.551,
+      "eval_steps_per_second": 0.502,
       "step": 400
     },
     {
       "epoch": 20.952380952380953,
-      "grad_norm": 296.1668395996094,
+      "grad_norm": 366.5032958984375,
       "learning_rate": 8.78e-06,
-      "loss": 108.6625,
+      "loss": 115.4496,
       "step": 440
     },
     {
       "epoch": 22.857142857142858,
-      "grad_norm": 270.3917541503906,
+      "grad_norm": 435.5931701660156,
       "learning_rate": 9.58e-06,
-      "loss": 86.1836,
+      "loss": 92.9752,
       "step": 480
     },
     {
       "epoch": 24.761904761904763,
-      "grad_norm": 252.61672973632812,
+      "grad_norm": 232.87234497070312,
       "learning_rate": 9.993604266900276e-06,
-      "loss": 66.8246,
+      "loss": 78.7805,
       "step": 520
     },
     {
       "epoch": 26.666666666666668,
-      "grad_norm": 176.43890380859375,
+      "grad_norm": 188.52627563476562,
       "learning_rate": 9.938441702975689e-06,
-      "loss": 52.6602,
+      "loss": 59.0797,
       "step": 560
     },
     {
       "epoch": 28.571428571428573,
-      "grad_norm": 143.43080139160156,
+      "grad_norm": 177.4958038330078,
       "learning_rate": 9.827324621255206e-06,
-      "loss": 45.2324,
+      "loss": 46.9344,
       "step": 600
     },
     {
       "epoch": 28.571428571428573,
-      "eval_loss": 131.0333709716797,
-      "eval_runtime": 7.5366,
-      "eval_samples_per_second": 15.392,
-      "eval_steps_per_second": 0.531,
+      "eval_loss": 129.46548461914062,
+      "eval_runtime": 7.2365,
+      "eval_samples_per_second": 16.03,
+      "eval_steps_per_second": 0.553,
       "step": 600
     },
     {
       "epoch": 30.476190476190474,
-      "grad_norm": 125.09796142578125,
+      "grad_norm": 135.09072875976562,
       "learning_rate": 9.661512022674852e-06,
-      "loss": 36.3571,
+      "loss": 39.1039,
       "step": 640
     },
     {
       "epoch": 32.38095238095238,
-      "grad_norm": 130.8366241455078,
+      "grad_norm": 128.56063842773438,
       "learning_rate": 9.4428826302319e-06,
-      "loss": 30.4001,
+      "loss": 32.5044,
       "step": 680
     },
     {
       "epoch": 34.285714285714285,
-      "grad_norm": 98.4572525024414,
+      "grad_norm": 99.94125366210938,
       "learning_rate": 9.173913602302594e-06,
-      "loss": 27.7692,
+      "loss": 27.9798,
       "step": 720
     },
     {
       "epoch": 36.19047619047619,
-      "grad_norm": 118.89614868164062,
+      "grad_norm": 128.35171508789062,
       "learning_rate": 8.857652465447947e-06,
-      "loss": 24.9675,
+      "loss": 24.9475,
       "step": 760
     },
     {
       "epoch": 38.095238095238095,
-      "grad_norm": 89.57968139648438,
+      "grad_norm": 113.99468231201172,
       "learning_rate": 8.497682584720065e-06,
-      "loss": 24.3789,
+      "loss": 22.38,
       "step": 800
     },
     {
       "epoch": 38.095238095238095,
-      "eval_loss": 80.99005126953125,
-      "eval_runtime": 7.2774,
-      "eval_samples_per_second": 15.94,
-      "eval_steps_per_second": 0.55,
+      "eval_loss": 75.64146423339844,
+      "eval_runtime": 7.294,
+      "eval_samples_per_second": 15.904,
+      "eval_steps_per_second": 0.548,
       "step": 800
     },
     {
       "epoch": 40.0,
-      "grad_norm": 132.4896240234375,
+      "grad_norm": 190.8120574951172,
       "learning_rate": 8.098082562704253e-06,
-      "loss": 20.1034,
+      "loss": 19.868,
       "step": 840
     },
     {
       "epoch": 41.904761904761905,
-      "grad_norm": 81.84835815429688,
+      "grad_norm": 58.219482421875,
       "learning_rate": 7.663380027322022e-06,
-      "loss": 18.1971,
+      "loss": 22.0343,
       "step": 880
     },
     {
       "epoch": 43.80952380952381,
-      "grad_norm": 55.926551818847656,
+      "grad_norm": 89.1216049194336,
       "learning_rate": 7.1985003319975696e-06,
-      "loss": 17.9396,
+      "loss": 16.9501,
       "step": 920
     },
     {
       "epoch": 45.714285714285715,
-      "grad_norm": 64.777587890625,
+      "grad_norm": 147.2601776123047,
       "learning_rate": 6.708710749435427e-06,
-      "loss": 15.3426,
+      "loss": 15.9704,
       "step": 960
     },
     {
       "epoch": 47.61904761904762,
-      "grad_norm": 53.17697525024414,
+      "grad_norm": 103.56001281738281,
       "learning_rate": 6.199560791316009e-06,
-      "loss": 16.3762,
+      "loss": 15.38,
       "step": 1000
     },
     {
       "epoch": 47.61904761904762,
-      "eval_loss": 65.57749938964844,
-      "eval_runtime": 7.054,
-      "eval_samples_per_second": 16.445,
-      "eval_steps_per_second": 0.567,
+      "eval_loss": 79.28771209716797,
+      "eval_runtime": 7.3262,
+      "eval_samples_per_second": 15.834,
+      "eval_steps_per_second": 0.546,
       "step": 1000
     }
   ],

checkpoint-1000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c0e178d29d8bd1005e0ebacb559c58b57ef56dd83b9d009bde5540a5128a1d4a
+oid sha256:ba4185c232729e9f26479512d5ae60057198316f129ed28036d583f86d686e96
 size 5777

checkpoint-1200/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4d408ce3ae88f5fc0fd2cdd76846c1a2f90f63dcb1ee9d633e77e90da0743c34
+oid sha256:d0cc801e4b2876cb018d78b072e584c108ec7af9da7588b7f7b9d75c799ba940
 size 306699044

checkpoint-1200/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a4e0185ae80fb21f14573e6ecee06fff8101e23938cb3ecdbca6303f3084883d
+oid sha256:1f60249dca57809c46c4992bac2974000714ff38516f20e2faa0c142a9841632
 size 611580898

checkpoint-1200/trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_global_step": 1200,
-  "best_metric": 58.122344970703125,
+  "best_metric": 64.76508331298828,
   "best_model_checkpoint": "rtdetr_voucher_model/checkpoint-1200",
   "epoch": 57.142857142857146,
   "eval_steps": 200,
@@ -11,260 +11,260 @@
   "log_history": [
     {
       "epoch": 1.9047619047619047,
-      "grad_norm": 10800.818359375,
+      "grad_norm": 23553.73828125,
       "learning_rate": 7.8e-07,
-      "loss": 1388.4377,
+      "loss": 1390.8067,
       "step": 40
     },
     {
       "epoch": 3.8095238095238093,
-      "grad_norm": 21733.138671875,
+      "grad_norm": 7427.158203125,
       "learning_rate": 1.5800000000000001e-06,
-      "loss": 1273.0859,
+      "loss": 1273.2542,
       "step": 80
     },
     {
       "epoch": 5.714285714285714,
-      "grad_norm": 2858.446044921875,
+      "grad_norm": 6636.22314453125,
       "learning_rate": 2.38e-06,
-      "loss": 1063.7887,
+      "loss": 1059.9522,
       "step": 120
     },
     {
       "epoch": 7.619047619047619,
-      "grad_norm": 4524.0263671875,
+      "grad_norm": 1760.4281005859375,
       "learning_rate": 3.1800000000000005e-06,
-      "loss": 771.9299,
+      "loss": 790.0425,
       "step": 160
     },
     {
       "epoch": 9.523809523809524,
-      "grad_norm": 2449.484130859375,
+      "grad_norm": 1625.7744140625,
       "learning_rate": 3.980000000000001e-06,
-      "loss": 519.636,
+      "loss": 517.1978,
       "step": 200
     },
     {
       "epoch": 9.523809523809524,
-      "eval_loss": 561.4364624023438,
-      "eval_runtime": 8.2677,
-      "eval_samples_per_second": 14.03,
-      "eval_steps_per_second": 0.484,
+      "eval_loss": 535.0709838867188,
+      "eval_runtime": 8.3595,
+      "eval_samples_per_second": 13.876,
+      "eval_steps_per_second": 0.478,
       "step": 200
     },
     {
       "epoch": 11.428571428571429,
-      "grad_norm": 3173.60546875,
+      "grad_norm": 1354.9310302734375,
       "learning_rate": 4.78e-06,
-      "loss": 396.2762,
+      "loss": 353.8703,
       "step": 240
     },
     {
       "epoch": 13.333333333333334,
-      "grad_norm": 957.3853149414062,
+      "grad_norm": 873.0985107421875,
       "learning_rate": 5.580000000000001e-06,
-      "loss": 317.4443,
+      "loss": 262.4233,
       "step": 280
     },
     {
       "epoch": 15.238095238095237,
-      "grad_norm": 2215.865966796875,
+      "grad_norm": 1516.567626953125,
       "learning_rate": 6.380000000000001e-06,
-      "loss": 244.5671,
+      "loss": 205.628,
       "step": 320
     },
     {
       "epoch": 17.142857142857142,
-      "grad_norm": 494.6834716796875,
+      "grad_norm": 5139.3779296875,
       "learning_rate": 7.180000000000001e-06,
-      "loss": 200.2045,
+      "loss": 174.0845,
       "step": 360
     },
     {
       "epoch": 19.047619047619047,
-      "grad_norm": 389.99395751953125,
+      "grad_norm": 480.6575622558594,
       "learning_rate": 7.980000000000002e-06,
-      "loss": 143.4672,
+      "loss": 142.4243,
       "step": 400
     },
     {
       "epoch": 19.047619047619047,
-      "eval_loss": 212.24693298339844,
-      "eval_runtime": 7.6261,
-      "eval_samples_per_second": 15.211,
-      "eval_steps_per_second": 0.525,
+      "eval_loss": 214.48715209960938,
+      "eval_runtime": 7.9717,
+      "eval_samples_per_second": 14.551,
+      "eval_steps_per_second": 0.502,
       "step": 400
     },
     {
       "epoch": 20.952380952380953,
-      "grad_norm": 296.1668395996094,
+      "grad_norm": 366.5032958984375,
       "learning_rate": 8.78e-06,
-      "loss": 108.6625,
+      "loss": 115.4496,
       "step": 440
     },
     {
       "epoch": 22.857142857142858,
-      "grad_norm": 270.3917541503906,
+      "grad_norm": 435.5931701660156,
       "learning_rate": 9.58e-06,
-      "loss": 86.1836,
+      "loss": 92.9752,
       "step": 480
     },
     {
       "epoch": 24.761904761904763,
-      "grad_norm": 252.61672973632812,
+      "grad_norm": 232.87234497070312,
       "learning_rate": 9.993604266900276e-06,
-      "loss": 66.8246,
+      "loss": 78.7805,
       "step": 520
     },
     {
       "epoch": 26.666666666666668,
-      "grad_norm": 176.43890380859375,
+      "grad_norm": 188.52627563476562,
       "learning_rate": 9.938441702975689e-06,
-      "loss": 52.6602,
+      "loss": 59.0797,
       "step": 560
     },
     {
       "epoch": 28.571428571428573,
-      "grad_norm": 143.43080139160156,
+      "grad_norm": 177.4958038330078,
       "learning_rate": 9.827324621255206e-06,
-      "loss": 45.2324,
+      "loss": 46.9344,
       "step": 600
     },
     {
       "epoch": 28.571428571428573,
-      "eval_loss": 131.0333709716797,
-      "eval_runtime": 7.5366,
-      "eval_samples_per_second": 15.392,
-      "eval_steps_per_second": 0.531,
+      "eval_loss": 129.46548461914062,
+      "eval_runtime": 7.2365,
+      "eval_samples_per_second": 16.03,
+      "eval_steps_per_second": 0.553,
       "step": 600
     },
     {
       "epoch": 30.476190476190474,
-      "grad_norm": 125.09796142578125,
+      "grad_norm": 135.09072875976562,
       "learning_rate": 9.661512022674852e-06,
-      "loss": 36.3571,
+      "loss": 39.1039,
       "step": 640
     },
     {
       "epoch": 32.38095238095238,
-      "grad_norm": 130.8366241455078,
+      "grad_norm": 128.56063842773438,
       "learning_rate": 9.4428826302319e-06,
-      "loss": 30.4001,
+      "loss": 32.5044,
       "step": 680
     },
     {
       "epoch": 34.285714285714285,
-      "grad_norm": 98.4572525024414,
+      "grad_norm": 99.94125366210938,
       "learning_rate": 9.173913602302594e-06,
-      "loss": 27.7692,
+      "loss": 27.9798,
       "step": 720
     },
     {
       "epoch": 36.19047619047619,
-      "grad_norm": 118.89614868164062,
+      "grad_norm": 128.35171508789062,
       "learning_rate": 8.857652465447947e-06,
-      "loss": 24.9675,
+      "loss": 24.9475,
       "step": 760
     },
     {
       "epoch": 38.095238095238095,
-      "grad_norm": 89.57968139648438,
+      "grad_norm": 113.99468231201172,
       "learning_rate": 8.497682584720065e-06,
-      "loss": 24.3789,
+      "loss": 22.38,
       "step": 800
     },
     {
       "epoch": 38.095238095238095,
-      "eval_loss": 80.99005126953125,
-      "eval_runtime": 7.2774,
-      "eval_samples_per_second": 15.94,
-      "eval_steps_per_second": 0.55,
+      "eval_loss": 75.64146423339844,
+      "eval_runtime": 7.294,
+      "eval_samples_per_second": 15.904,
+      "eval_steps_per_second": 0.548,
       "step": 800
     },
     {
       "epoch": 40.0,
-      "grad_norm": 132.4896240234375,
+      "grad_norm": 190.8120574951172,
       "learning_rate": 8.098082562704253e-06,
-      "loss": 20.1034,
+      "loss": 19.868,
       "step": 840
     },
     {
       "epoch": 41.904761904761905,
-      "grad_norm": 81.84835815429688,
+      "grad_norm": 58.219482421875,
       "learning_rate": 7.663380027322022e-06,
-      "loss": 18.1971,
+      "loss": 22.0343,
       "step": 880
     },
     {
       "epoch": 43.80952380952381,
-      "grad_norm": 55.926551818847656,
+      "grad_norm": 89.1216049194336,
       "learning_rate": 7.1985003319975696e-06,
-      "loss": 17.9396,
+      "loss": 16.9501,
       "step": 920
     },
     {
       "epoch": 45.714285714285715,
-      "grad_norm": 64.777587890625,
+      "grad_norm": 147.2601776123047,
       "learning_rate": 6.708710749435427e-06,
-      "loss": 15.3426,
+      "loss": 15.9704,
       "step": 960
     },
     {
       "epoch": 47.61904761904762,
-      "grad_norm": 53.17697525024414,
+      "grad_norm": 103.56001281738281,
       "learning_rate": 6.199560791316009e-06,
-      "loss": 16.3762,
+      "loss": 15.38,
       "step": 1000
     },
     {
       "epoch": 47.61904761904762,
-      "eval_loss": 65.57749938964844,
-      "eval_runtime": 7.054,
-      "eval_samples_per_second": 16.445,
-      "eval_steps_per_second": 0.567,
+      "eval_loss": 79.28771209716797,
+      "eval_runtime": 7.3262,
+      "eval_samples_per_second": 15.834,
+      "eval_steps_per_second": 0.546,
       "step": 1000
     },
     {
       "epoch": 49.523809523809526,
-      "grad_norm": 140.15521240234375,
+      "grad_norm": 49.25249099731445,
       "learning_rate": 5.676819330110808e-06,
-      "loss": 13.8692,
+      "loss": 14.5406,
       "step": 1040
     },
     {
       "epoch": 51.42857142857143,
-      "grad_norm": 61.619049072265625,
+      "grad_norm": 116.98213195800781,
       "learning_rate": 5.14640923545217e-06,
-      "loss": 15.6367,
+      "loss": 14.0173,
       "step": 1080
     },
     {
       "epoch": 53.333333333333336,
-      "grad_norm": 147.7701873779297,
+      "grad_norm": 72.93340301513672,
       "learning_rate": 4.614340265653695e-06,
-      "loss": 13.0418,
+      "loss": 13.7052,
       "step": 1120
     },
     {
       "epoch": 55.23809523809524,
-      "grad_norm": 77.45187377929688,
+      "grad_norm": 105.87384796142578,
       "learning_rate": 4.086640974747176e-06,
-      "loss": 12.7083,
+      "loss": 15.4002,
       "step": 1160
     },
     {
       "epoch": 57.142857142857146,
-      "grad_norm": 215.86756896972656,
+      "grad_norm": 59.23909378051758,
       "learning_rate": 3.569290406556519e-06,
-      "loss": 13.65,
+      "loss": 14.1171,
       "step": 1200
     },
     {
       "epoch": 57.142857142857146,
-      "eval_loss": 58.122344970703125,
-      "eval_runtime": 7.0277,
-      "eval_samples_per_second": 16.506,
-      "eval_steps_per_second": 0.569,
+      "eval_loss": 64.76508331298828,
+      "eval_runtime": 7.1521,
+      "eval_samples_per_second": 16.219,
+      "eval_steps_per_second": 0.559,
       "step": 1200
     }
   ],

checkpoint-1200/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c0e178d29d8bd1005e0ebacb559c58b57ef56dd83b9d009bde5540a5128a1d4a
+oid sha256:ba4185c232729e9f26479512d5ae60057198316f129ed28036d583f86d686e96
 size 5777

checkpoint-1400/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9857953e8c6751ca6b9433ec63d899eda71d75aad98c054ee0415c1dc3b37be2
+oid sha256:04b94f7cad04f0747f4fc31287436697a425f647e5aefc95be51e2c7a837b4c9
 size 306699044

checkpoint-1400/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:af83e894e9db3ea3ef841289f868d2ae7b8734b566076e82fec88fb48d45a921
+oid sha256:a5b3060dd15542402c6e8fb677b15552983f39dcb3f4878c90473c7cacf46766
 size 611580898

checkpoint-1400/trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_global_step": 1200,
-  "best_metric": 58.122344970703125,
+  "best_metric": 64.76508331298828,
   "best_model_checkpoint": "rtdetr_voucher_model/checkpoint-1200",
   "epoch": 66.66666666666667,
   "eval_steps": 200,
@@ -11,303 +11,303 @@
   "log_history": [
     {
       "epoch": 1.9047619047619047,
-      "grad_norm": 10800.818359375,
+      "grad_norm": 23553.73828125,
       "learning_rate": 7.8e-07,
-      "loss": 1388.4377,
+      "loss": 1390.8067,
       "step": 40
     },
     {
       "epoch": 3.8095238095238093,
-      "grad_norm": 21733.138671875,
+      "grad_norm": 7427.158203125,
       "learning_rate": 1.5800000000000001e-06,
-      "loss": 1273.0859,
+      "loss": 1273.2542,
       "step": 80
     },
     {
       "epoch": 5.714285714285714,
-      "grad_norm": 2858.446044921875,
+      "grad_norm": 6636.22314453125,
       "learning_rate": 2.38e-06,
-      "loss": 1063.7887,
+      "loss": 1059.9522,
       "step": 120
     },
     {
       "epoch": 7.619047619047619,
-      "grad_norm": 4524.0263671875,
+      "grad_norm": 1760.4281005859375,
       "learning_rate": 3.1800000000000005e-06,
-      "loss": 771.9299,
+      "loss": 790.0425,
       "step": 160
     },
     {
       "epoch": 9.523809523809524,
-      "grad_norm": 2449.484130859375,
+      "grad_norm": 1625.7744140625,
       "learning_rate": 3.980000000000001e-06,
-      "loss": 519.636,
+      "loss": 517.1978,
       "step": 200
     },
     {
       "epoch": 9.523809523809524,
-      "eval_loss": 561.4364624023438,
-      "eval_runtime": 8.2677,
-      "eval_samples_per_second": 14.03,
-      "eval_steps_per_second": 0.484,
+      "eval_loss": 535.0709838867188,
+      "eval_runtime": 8.3595,
+      "eval_samples_per_second": 13.876,
+      "eval_steps_per_second": 0.478,
       "step": 200
     },
     {
       "epoch": 11.428571428571429,
-      "grad_norm": 3173.60546875,
+      "grad_norm": 1354.9310302734375,
       "learning_rate": 4.78e-06,
-      "loss": 396.2762,
+      "loss": 353.8703,
       "step": 240
     },
     {
       "epoch": 13.333333333333334,
-      "grad_norm": 957.3853149414062,
+      "grad_norm": 873.0985107421875,
       "learning_rate": 5.580000000000001e-06,
-      "loss": 317.4443,
+      "loss": 262.4233,
       "step": 280
     },
     {
       "epoch": 15.238095238095237,
-      "grad_norm": 2215.865966796875,
+      "grad_norm": 1516.567626953125,
       "learning_rate": 6.380000000000001e-06,
-      "loss": 244.5671,
+      "loss": 205.628,
       "step": 320
     },
     {
       "epoch": 17.142857142857142,
-      "grad_norm": 494.6834716796875,
+      "grad_norm": 5139.3779296875,
       "learning_rate": 7.180000000000001e-06,
-      "loss": 200.2045,
+      "loss": 174.0845,
       "step": 360
     },
     {
       "epoch": 19.047619047619047,
-      "grad_norm": 389.99395751953125,
+      "grad_norm": 480.6575622558594,
       "learning_rate": 7.980000000000002e-06,
-      "loss": 143.4672,
+      "loss": 142.4243,
       "step": 400
     },
     {
       "epoch": 19.047619047619047,
-      "eval_loss": 212.24693298339844,
-      "eval_runtime": 7.6261,
-      "eval_samples_per_second": 15.211,
-      "eval_steps_per_second": 0.525,
+      "eval_loss": 214.48715209960938,
+      "eval_runtime": 7.9717,
+      "eval_samples_per_second": 14.551,
+      "eval_steps_per_second": 0.502,
       "step": 400
     },
     {
       "epoch": 20.952380952380953,
-      "grad_norm": 296.1668395996094,
+      "grad_norm": 366.5032958984375,
       "learning_rate": 8.78e-06,
-      "loss": 108.6625,
+      "loss": 115.4496,
       "step": 440
     },
     {
       "epoch": 22.857142857142858,
-      "grad_norm": 270.3917541503906,
+      "grad_norm": 435.5931701660156,
       "learning_rate": 9.58e-06,
-      "loss": 86.1836,
+      "loss": 92.9752,
       "step": 480
     },
     {
       "epoch": 24.761904761904763,
-      "grad_norm": 252.61672973632812,
+      "grad_norm": 232.87234497070312,
       "learning_rate": 9.993604266900276e-06,
-      "loss": 66.8246,
+      "loss": 78.7805,
       "step": 520
     },
     {
       "epoch": 26.666666666666668,
-      "grad_norm": 176.43890380859375,
+      "grad_norm": 188.52627563476562,
       "learning_rate": 9.938441702975689e-06,
-      "loss": 52.6602,
+      "loss": 59.0797,
       "step": 560
     },
     {
       "epoch": 28.571428571428573,
-      "grad_norm": 143.43080139160156,
+      "grad_norm": 177.4958038330078,
       "learning_rate": 9.827324621255206e-06,
-      "loss": 45.2324,
+      "loss": 46.9344,
       "step": 600
     },
     {
       "epoch": 28.571428571428573,
-      "eval_loss": 131.0333709716797,
-      "eval_runtime": 7.5366,
-      "eval_samples_per_second": 15.392,
-      "eval_steps_per_second": 0.531,
+      "eval_loss": 129.46548461914062,
+      "eval_runtime": 7.2365,
+      "eval_samples_per_second": 16.03,
+      "eval_steps_per_second": 0.553,
       "step": 600
     },
     {
       "epoch": 30.476190476190474,
-      "grad_norm": 125.09796142578125,
+      "grad_norm": 135.09072875976562,
       "learning_rate": 9.661512022674852e-06,
-      "loss": 36.3571,
+      "loss": 39.1039,
       "step": 640
     },
     {
       "epoch": 32.38095238095238,
-      "grad_norm": 130.8366241455078,
+      "grad_norm": 128.56063842773438,
       "learning_rate": 9.4428826302319e-06,
-      "loss": 30.4001,
+      "loss": 32.5044,
       "step": 680
     },
     {
       "epoch": 34.285714285714285,
-      "grad_norm": 98.4572525024414,
+      "grad_norm": 99.94125366210938,
       "learning_rate": 9.173913602302594e-06,
-      "loss": 27.7692,
+      "loss": 27.9798,
       "step": 720
     },
     {
       "epoch": 36.19047619047619,
-      "grad_norm": 118.89614868164062,
+      "grad_norm": 128.35171508789062,
       "learning_rate": 8.857652465447947e-06,
-      "loss": 24.9675,
+      "loss": 24.9475,
       "step": 760
     },
     {
       "epoch": 38.095238095238095,
-      "grad_norm": 89.57968139648438,
+      "grad_norm": 113.99468231201172,
       "learning_rate": 8.497682584720065e-06,
-      "loss": 24.3789,
+      "loss": 22.38,
       "step": 800
     },
     {
       "epoch": 38.095238095238095,
-      "eval_loss": 80.99005126953125,
-      "eval_runtime": 7.2774,
-      "eval_samples_per_second": 15.94,
-      "eval_steps_per_second": 0.55,
+      "eval_loss": 75.64146423339844,
+      "eval_runtime": 7.294,
+      "eval_samples_per_second": 15.904,
+      "eval_steps_per_second": 0.548,
       "step": 800
     },
     {
       "epoch": 40.0,
-      "grad_norm": 132.4896240234375,
+      "grad_norm": 190.8120574951172,
       "learning_rate": 8.098082562704253e-06,
-      "loss": 20.1034,
+      "loss": 19.868,
       "step": 840
     },
     {
       "epoch": 41.904761904761905,
-      "grad_norm": 81.84835815429688,
+      "grad_norm": 58.219482421875,
       "learning_rate": 7.663380027322022e-06,
-      "loss": 18.1971,
+      "loss": 22.0343,
       "step": 880
     },
     {
       "epoch": 43.80952380952381,
-      "grad_norm": 55.926551818847656,
+      "grad_norm": 89.1216049194336,
       "learning_rate": 7.1985003319975696e-06,
-      "loss": 17.9396,
+      "loss": 16.9501,
       "step": 920
     },
     {
       "epoch": 45.714285714285715,
-      "grad_norm": 64.777587890625,
+      "grad_norm": 147.2601776123047,
       "learning_rate": 6.708710749435427e-06,
-      "loss": 15.3426,
+      "loss": 15.9704,
       "step": 960
     },
     {
       "epoch": 47.61904761904762,
-      "grad_norm": 53.17697525024414,
+      "grad_norm": 103.56001281738281,
       "learning_rate": 6.199560791316009e-06,
-      "loss": 16.3762,
+      "loss": 15.38,
       "step": 1000
     },
     {
       "epoch": 47.61904761904762,
-      "eval_loss": 65.57749938964844,
-      "eval_runtime": 7.054,
-      "eval_samples_per_second": 16.445,
-      "eval_steps_per_second": 0.567,
+      "eval_loss": 79.28771209716797,
+      "eval_runtime": 7.3262,
+      "eval_samples_per_second": 15.834,
+      "eval_steps_per_second": 0.546,
       "step": 1000
     },
     {
       "epoch": 49.523809523809526,
-      "grad_norm": 140.15521240234375,
+      "grad_norm": 49.25249099731445,
       "learning_rate": 5.676819330110808e-06,
-      "loss": 13.8692,
+      "loss": 14.5406,
       "step": 1040
     },
     {
       "epoch": 51.42857142857143,
-      "grad_norm": 61.619049072265625,
+      "grad_norm": 116.98213195800781,
       "learning_rate": 5.14640923545217e-06,
-      "loss": 15.6367,
+      "loss": 14.0173,
       "step": 1080
     },
     {
       "epoch": 53.333333333333336,
-      "grad_norm": 147.7701873779297,
+      "grad_norm": 72.93340301513672,
       "learning_rate": 4.614340265653695e-06,
-      "loss": 13.0418,
+      "loss": 13.7052,
       "step": 1120
     },
     {
       "epoch": 55.23809523809524,
-      "grad_norm": 77.45187377929688,
+      "grad_norm": 105.87384796142578,
       "learning_rate": 4.086640974747176e-06,
-      "loss": 12.7083,
+      "loss": 15.4002,
       "step": 1160
     },
     {
       "epoch": 57.142857142857146,
-      "grad_norm": 215.86756896972656,
+      "grad_norm": 59.23909378051758,
       "learning_rate": 3.569290406556519e-06,
-      "loss": 13.65,
+      "loss": 14.1171,
       "step": 1200
     },
     {
       "epoch": 57.142857142857146,
-      "eval_loss": 58.122344970703125,
-      "eval_runtime": 7.0277,
-      "eval_samples_per_second": 16.506,
-      "eval_steps_per_second": 0.569,
+      "eval_loss": 64.76508331298828,
+      "eval_runtime": 7.1521,
+      "eval_samples_per_second": 16.219,
+      "eval_steps_per_second": 0.559,
       "step": 1200
     },
     {
       "epoch": 59.04761904761905,
-      "grad_norm": 76.79309844970703,
+      "grad_norm": 98.89187622070312,
       "learning_rate": 3.068150349742125e-06,
-      "loss": 11.9538,
+      "loss": 12.3597,
       "step": 1240
     },
     {
       "epoch": 60.95238095238095,
-      "grad_norm": 52.94865417480469,
+      "grad_norm": 179.5811004638672,
       "learning_rate": 2.5888989213931047e-06,
-      "loss": 11.5392,
+      "loss": 12.125,
       "step": 1280
     },
     {
       "epoch": 62.857142857142854,
-      "grad_norm": 153.39662170410156,
+      "grad_norm": 41.74519729614258,
       "learning_rate": 2.1369662316918095e-06,
-      "loss": 12.1913,
+      "loss": 11.7567,
       "step": 1320
     },
     {
       "epoch": 64.76190476190476,
-      "grad_norm": 139.7332305908203,
+      "grad_norm": 60.97040557861328,
       "learning_rate": 1.717472858595694e-06,
-      "loss": 11.2592,
+      "loss": 12.3442,
       "step": 1360
     },
     {
       "epoch": 66.66666666666667,
-      "grad_norm": 39.22611999511719,
+      "grad_norm": 197.75689697265625,
       "learning_rate": 1.335171829642984e-06,
-      "loss": 11.1986,
+      "loss": 11.2475,
       "step": 1400
     },
     {
       "epoch": 66.66666666666667,
-      "eval_loss": 60.4874382019043,
-      "eval_runtime": 7.0545,
-      "eval_samples_per_second": 16.443,
-      "eval_steps_per_second": 0.567,
+      "eval_loss": 67.15937805175781,
+      "eval_runtime": 7.2137,
+      "eval_samples_per_second": 16.08,
+      "eval_steps_per_second": 0.554,
       "step": 1400
     }
   ],

checkpoint-1400/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c0e178d29d8bd1005e0ebacb559c58b57ef56dd83b9d009bde5540a5128a1d4a
+oid sha256:ba4185c232729e9f26479512d5ae60057198316f129ed28036d583f86d686e96
 size 5777

checkpoint-1600/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8fc83e9ffb9854d6f50d18b71d3610cb9b411e48e241f6e96354f2fd615e20c4
+oid sha256:26f7b879ae25a9c83fa3e849fdb085366c9c812655437eeedbdf446b895a6cc5
 size 306699044

checkpoint-1600/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:eeced65cec309962d348ad3cc1ce6362b5a5185260bc0875e65a5f20507842a8
+oid sha256:42bd1f3c13814d5212492c77c1c29fe168190fb0ed3c50250e6af351070f0d2b
 size 611580898

checkpoint-1600/trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_global_step": 1600,
-  "best_metric": 56.9376335144043,
+  "best_metric": 62.760154724121094,
   "best_model_checkpoint": "rtdetr_voucher_model/checkpoint-1600",
   "epoch": 76.19047619047619,
   "eval_steps": 200,
@@ -11,346 +11,346 @@
   "log_history": [
     {
       "epoch": 1.9047619047619047,
-      "grad_norm": 10800.818359375,
+      "grad_norm": 23553.73828125,
       "learning_rate": 7.8e-07,
-      "loss": 1388.4377,
+      "loss": 1390.8067,
       "step": 40
     },
     {
       "epoch": 3.8095238095238093,
-      "grad_norm": 21733.138671875,
+      "grad_norm": 7427.158203125,
       "learning_rate": 1.5800000000000001e-06,
-      "loss": 1273.0859,
+      "loss": 1273.2542,
       "step": 80
     },
     {
       "epoch": 5.714285714285714,
-      "grad_norm": 2858.446044921875,
+      "grad_norm": 6636.22314453125,
       "learning_rate": 2.38e-06,
-      "loss": 1063.7887,
+      "loss": 1059.9522,
       "step": 120
     },
     {
       "epoch": 7.619047619047619,
-      "grad_norm": 4524.0263671875,
+      "grad_norm": 1760.4281005859375,
       "learning_rate": 3.1800000000000005e-06,
-      "loss": 771.9299,
+      "loss": 790.0425,
       "step": 160
     },
     {
       "epoch": 9.523809523809524,
-      "grad_norm": 2449.484130859375,
+      "grad_norm": 1625.7744140625,
       "learning_rate": 3.980000000000001e-06,
-      "loss": 519.636,
+      "loss": 517.1978,
       "step": 200
     },
     {
       "epoch": 9.523809523809524,
-      "eval_loss": 561.4364624023438,
-      "eval_runtime": 8.2677,
-      "eval_samples_per_second": 14.03,
-      "eval_steps_per_second": 0.484,
+      "eval_loss": 535.0709838867188,
+      "eval_runtime": 8.3595,
+      "eval_samples_per_second": 13.876,
+      "eval_steps_per_second": 0.478,
       "step": 200
     },
     {
       "epoch": 11.428571428571429,
-      "grad_norm": 3173.60546875,
+      "grad_norm": 1354.9310302734375,
       "learning_rate": 4.78e-06,
-      "loss": 396.2762,
+      "loss": 353.8703,
       "step": 240
     },
     {
       "epoch": 13.333333333333334,
-      "grad_norm": 957.3853149414062,
+      "grad_norm": 873.0985107421875,
       "learning_rate": 5.580000000000001e-06,
-      "loss": 317.4443,
+      "loss": 262.4233,
       "step": 280
     },
     {
       "epoch": 15.238095238095237,
-      "grad_norm": 2215.865966796875,
+      "grad_norm": 1516.567626953125,
       "learning_rate": 6.380000000000001e-06,
-      "loss": 244.5671,
+      "loss": 205.628,
       "step": 320
     },
     {
       "epoch": 17.142857142857142,
-      "grad_norm": 494.6834716796875,
+      "grad_norm": 5139.3779296875,
       "learning_rate": 7.180000000000001e-06,
-      "loss": 200.2045,
+      "loss": 174.0845,
       "step": 360
     },
     {
       "epoch": 19.047619047619047,
-      "grad_norm": 389.99395751953125,
+      "grad_norm": 480.6575622558594,
       "learning_rate": 7.980000000000002e-06,
-      "loss": 143.4672,
+      "loss": 142.4243,
       "step": 400
     },
     {
       "epoch": 19.047619047619047,
-      "eval_loss": 212.24693298339844,
-      "eval_runtime": 7.6261,
-      "eval_samples_per_second": 15.211,
-      "eval_steps_per_second": 0.525,
+      "eval_loss": 214.48715209960938,
+      "eval_runtime": 7.9717,
+      "eval_samples_per_second": 14.551,
+      "eval_steps_per_second": 0.502,
       "step": 400
     },
     {
       "epoch": 20.952380952380953,
-      "grad_norm": 296.1668395996094,
+      "grad_norm": 366.5032958984375,
       "learning_rate": 8.78e-06,
-      "loss": 108.6625,
+      "loss": 115.4496,
       "step": 440
     },
     {
       "epoch": 22.857142857142858,
-      "grad_norm": 270.3917541503906,
+      "grad_norm": 435.5931701660156,
       "learning_rate": 9.58e-06,
-      "loss": 86.1836,
+      "loss": 92.9752,
       "step": 480
     },
     {
       "epoch": 24.761904761904763,
-      "grad_norm": 252.61672973632812,
+      "grad_norm": 232.87234497070312,
       "learning_rate": 9.993604266900276e-06,
-      "loss": 66.8246,
+      "loss": 78.7805,
       "step": 520
     },
     {
       "epoch": 26.666666666666668,
-      "grad_norm": 176.43890380859375,
+      "grad_norm": 188.52627563476562,
       "learning_rate": 9.938441702975689e-06,
-      "loss": 52.6602,
+      "loss": 59.0797,
       "step": 560
     },
     {
       "epoch": 28.571428571428573,
-      "grad_norm": 143.43080139160156,
+      "grad_norm": 177.4958038330078,
       "learning_rate": 9.827324621255206e-06,
-      "loss": 45.2324,
+      "loss": 46.9344,
       "step": 600
     },
     {
       "epoch": 28.571428571428573,
-      "eval_loss": 131.0333709716797,
-      "eval_runtime": 7.5366,
-      "eval_samples_per_second": 15.392,
-      "eval_steps_per_second": 0.531,
+      "eval_loss": 129.46548461914062,
+      "eval_runtime": 7.2365,
+      "eval_samples_per_second": 16.03,
+      "eval_steps_per_second": 0.553,
       "step": 600
     },
     {
       "epoch": 30.476190476190474,
-      "grad_norm": 125.09796142578125,
+      "grad_norm": 135.09072875976562,
       "learning_rate": 9.661512022674852e-06,
-      "loss": 36.3571,
+      "loss": 39.1039,
       "step": 640
     },
     {
       "epoch": 32.38095238095238,
-      "grad_norm": 130.8366241455078,
+      "grad_norm": 128.56063842773438,
       "learning_rate": 9.4428826302319e-06,
-      "loss": 30.4001,
+      "loss": 32.5044,
       "step": 680
     },
     {
       "epoch": 34.285714285714285,
-      "grad_norm": 98.4572525024414,
+      "grad_norm": 99.94125366210938,
       "learning_rate": 9.173913602302594e-06,
-      "loss": 27.7692,
+      "loss": 27.9798,
       "step": 720
     },
     {
       "epoch": 36.19047619047619,
-      "grad_norm": 118.89614868164062,
+      "grad_norm": 128.35171508789062,
       "learning_rate": 8.857652465447947e-06,
-      "loss": 24.9675,
+      "loss": 24.9475,
       "step": 760
     },
     {
       "epoch": 38.095238095238095,
-      "grad_norm": 89.57968139648438,
+      "grad_norm": 113.99468231201172,
       "learning_rate": 8.497682584720065e-06,
-      "loss": 24.3789,
+      "loss": 22.38,
       "step": 800
     },
     {
       "epoch": 38.095238095238095,
-      "eval_loss": 80.99005126953125,
-      "eval_runtime": 7.2774,
-      "eval_samples_per_second": 15.94,
-      "eval_steps_per_second": 0.55,
+      "eval_loss": 75.64146423339844,
+      "eval_runtime": 7.294,
+      "eval_samples_per_second": 15.904,
+      "eval_steps_per_second": 0.548,
       "step": 800
     },
     {
       "epoch": 40.0,
-      "grad_norm": 132.4896240234375,
+      "grad_norm": 190.8120574951172,
       "learning_rate": 8.098082562704253e-06,
-      "loss": 20.1034,
+      "loss": 19.868,
       "step": 840
     },
     {
       "epoch": 41.904761904761905,
-      "grad_norm": 81.84835815429688,
+      "grad_norm": 58.219482421875,
       "learning_rate": 7.663380027322022e-06,
-      "loss": 18.1971,
+      "loss": 22.0343,
       "step": 880
     },
     {
       "epoch": 43.80952380952381,
-      "grad_norm": 55.926551818847656,
+      "grad_norm": 89.1216049194336,
       "learning_rate": 7.1985003319975696e-06,
-      "loss": 17.9396,
+      "loss": 16.9501,
       "step": 920
     },
     {
       "epoch": 45.714285714285715,
-      "grad_norm": 64.777587890625,
+      "grad_norm": 147.2601776123047,
       "learning_rate": 6.708710749435427e-06,
-      "loss": 15.3426,
+      "loss": 15.9704,
       "step": 960
     },
     {
       "epoch": 47.61904761904762,
-      "grad_norm": 53.17697525024414,
+      "grad_norm": 103.56001281738281,
       "learning_rate": 6.199560791316009e-06,
-      "loss": 16.3762,
+      "loss": 15.38,
       "step": 1000
     },
     {
       "epoch": 47.61904761904762,
-      "eval_loss": 65.57749938964844,
-      "eval_runtime": 7.054,
-      "eval_samples_per_second": 16.445,
-      "eval_steps_per_second": 0.567,
+      "eval_loss": 79.28771209716797,
+      "eval_runtime": 7.3262,
+      "eval_samples_per_second": 15.834,
+      "eval_steps_per_second": 0.546,
       "step": 1000
     },
     {
       "epoch": 49.523809523809526,
-      "grad_norm": 140.15521240234375,
+      "grad_norm": 49.25249099731445,
       "learning_rate": 5.676819330110808e-06,
-      "loss": 13.8692,
+      "loss": 14.5406,
       "step": 1040
     },
     {
       "epoch": 51.42857142857143,
-      "grad_norm": 61.619049072265625,
+      "grad_norm": 116.98213195800781,
       "learning_rate": 5.14640923545217e-06,
-      "loss": 15.6367,
+      "loss": 14.0173,
       "step": 1080
     },
     {
       "epoch": 53.333333333333336,
-      "grad_norm": 147.7701873779297,
+      "grad_norm": 72.93340301513672,
       "learning_rate": 4.614340265653695e-06,
-      "loss": 13.0418,
+      "loss": 13.7052,
       "step": 1120
     },
     {
       "epoch": 55.23809523809524,
-      "grad_norm": 77.45187377929688,
+      "grad_norm": 105.87384796142578,
       "learning_rate": 4.086640974747176e-06,
-      "loss": 12.7083,
+      "loss": 15.4002,
       "step": 1160
     },
     {
       "epoch": 57.142857142857146,
-      "grad_norm": 215.86756896972656,
+      "grad_norm": 59.23909378051758,
       "learning_rate": 3.569290406556519e-06,
-      "loss": 13.65,
+      "loss": 14.1171,
       "step": 1200
     },
     {
       "epoch": 57.142857142857146,
-      "eval_loss": 58.122344970703125,
-      "eval_runtime": 7.0277,
-      "eval_samples_per_second": 16.506,
-      "eval_steps_per_second": 0.569,
+      "eval_loss": 64.76508331298828,
+      "eval_runtime": 7.1521,
+      "eval_samples_per_second": 16.219,
+      "eval_steps_per_second": 0.559,
       "step": 1200
     },
     {
       "epoch": 59.04761904761905,
-      "grad_norm": 76.79309844970703,
+      "grad_norm": 98.89187622070312,
       "learning_rate": 3.068150349742125e-06,
-      "loss": 11.9538,
+      "loss": 12.3597,
       "step": 1240
     },
     {
       "epoch": 60.95238095238095,
-      "grad_norm": 52.94865417480469,
+      "grad_norm": 179.5811004638672,
       "learning_rate": 2.5888989213931047e-06,
-      "loss": 11.5392,
+      "loss": 12.125,
       "step": 1280
     },
     {
       "epoch": 62.857142857142854,
-      "grad_norm": 153.39662170410156,
+      "grad_norm": 41.74519729614258,
       "learning_rate": 2.1369662316918095e-06,
-      "loss": 12.1913,
+      "loss": 11.7567,
       "step": 1320
     },
     {
       "epoch": 64.76190476190476,
-      "grad_norm": 139.7332305908203,
+      "grad_norm": 60.97040557861328,
       "learning_rate": 1.717472858595694e-06,
-      "loss": 11.2592,
+      "loss": 12.3442,
       "step": 1360
     },
     {
       "epoch": 66.66666666666667,
-      "grad_norm": 39.22611999511719,
+      "grad_norm": 197.75689697265625,
       "learning_rate": 1.335171829642984e-06,
-      "loss": 11.1986,
+      "loss": 11.2475,
       "step": 1400
     },
     {
       "epoch": 66.66666666666667,
-      "eval_loss": 60.4874382019043,
-      "eval_runtime": 7.0545,
-      "eval_samples_per_second": 16.443,
-      "eval_steps_per_second": 0.567,
+      "eval_loss": 67.15937805175781,
+      "eval_runtime": 7.2137,
+      "eval_samples_per_second": 16.08,
+      "eval_steps_per_second": 0.554,
       "step": 1400
     },
     {
       "epoch": 68.57142857142857,
-      "grad_norm": 80.5909423828125,
+      "grad_norm": 44.10291290283203,
       "learning_rate": 9.943947682514538e-07,
-      "loss": 11.4891,
+      "loss": 11.0752,
       "step": 1440
     },
     {
       "epoch": 70.47619047619048,
-      "grad_norm": 92.74817657470703,
+      "grad_norm": 54.24775314331055,
       "learning_rate": 6.990028146942752e-07,
-      "loss": 11.4975,
+      "loss": 10.9762,
       "step": 1480
     },
     {
       "epoch": 72.38095238095238,
-      "grad_norm": 58.092369079589844,
+      "grad_norm": 41.610679626464844,
       "learning_rate": 4.523428778378852e-07,
-      "loss": 11.2603,
+      "loss": 10.9561,
       "step": 1520
     },
     {
       "epoch": 74.28571428571429,
-      "grad_norm": 171.13815307617188,
+      "grad_norm": 109.53748321533203,
       "learning_rate": 2.5720971332717783e-07,
-      "loss": 10.5585,
+      "loss": 10.7585,
       "step": 1560
     },
     {
       "epoch": 76.19047619047619,
-      "grad_norm": 49.394981384277344,
+      "grad_norm": 49.795623779296875,
       "learning_rate": 1.158142578873156e-07,
-      "loss": 10.9979,
+      "loss": 11.5847,
       "step": 1600
     },
     {
       "epoch": 76.19047619047619,
-      "eval_loss": 56.9376335144043,
-      "eval_runtime": 7.1981,
-      "eval_samples_per_second": 16.115,
-      "eval_steps_per_second": 0.556,
+      "eval_loss": 62.760154724121094,
+      "eval_runtime": 7.2145,
+      "eval_samples_per_second": 16.079,
+      "eval_steps_per_second": 0.554,
       "step": 1600
     }
   ],

checkpoint-1600/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c0e178d29d8bd1005e0ebacb559c58b57ef56dd83b9d009bde5540a5128a1d4a
+oid sha256:ba4185c232729e9f26479512d5ae60057198316f129ed28036d583f86d686e96
 size 5777

checkpoint-1680/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2e6958fd8fe84e6136b94d989b92cd00c3dfc3947d99e2b30a1a4201250d271a
+oid sha256:e88b3beaaa4ee5fccf6c71430bbd8100000f005578591d22ab5699386131e4bc
 size 306699044

checkpoint-1680/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a82a34a038f51ea8dedcb9407415208265fc33837121a7bed52250968ca50771
+oid sha256:536085da2f334c9892cc0fd076775e648bdc3fc281e1b08ea6bcd728d51b725c
 size 611580898

checkpoint-1680/trainer_state.json

@@ -1,6 +1,6 @@
 {
   "best_global_step": 1600,
-  "best_metric": 56.9376335144043,
+  "best_metric": 62.760154724121094,
   "best_model_checkpoint": "rtdetr_voucher_model/checkpoint-1600",
   "epoch": 80.0,
   "eval_steps": 200,
@@ -11,360 +11,360 @@
   "log_history": [
     {
       "epoch": 1.9047619047619047,
-      "grad_norm": 10800.818359375,
+      "grad_norm": 23553.73828125,
       "learning_rate": 7.8e-07,
-      "loss": 1388.4377,
+      "loss": 1390.8067,
       "step": 40
     },
     {
       "epoch": 3.8095238095238093,
-      "grad_norm": 21733.138671875,
+      "grad_norm": 7427.158203125,
       "learning_rate": 1.5800000000000001e-06,
-      "loss": 1273.0859,
+      "loss": 1273.2542,
       "step": 80
     },
     {
       "epoch": 5.714285714285714,
-      "grad_norm": 2858.446044921875,
+      "grad_norm": 6636.22314453125,
       "learning_rate": 2.38e-06,
-      "loss": 1063.7887,
+      "loss": 1059.9522,
       "step": 120
     },
     {
       "epoch": 7.619047619047619,
-      "grad_norm": 4524.0263671875,
+      "grad_norm": 1760.4281005859375,
       "learning_rate": 3.1800000000000005e-06,
-      "loss": 771.9299,
+      "loss": 790.0425,
       "step": 160
     },
     {
       "epoch": 9.523809523809524,
-      "grad_norm": 2449.484130859375,
+      "grad_norm": 1625.7744140625,
       "learning_rate": 3.980000000000001e-06,
-      "loss": 519.636,
+      "loss": 517.1978,
       "step": 200
     },
     {
       "epoch": 9.523809523809524,
-      "eval_loss": 561.4364624023438,
-      "eval_runtime": 8.2677,
-      "eval_samples_per_second": 14.03,
-      "eval_steps_per_second": 0.484,
+      "eval_loss": 535.0709838867188,
+      "eval_runtime": 8.3595,
+      "eval_samples_per_second": 13.876,
+      "eval_steps_per_second": 0.478,
       "step": 200
     },
     {
       "epoch": 11.428571428571429,
-      "grad_norm": 3173.60546875,
+      "grad_norm": 1354.9310302734375,
       "learning_rate": 4.78e-06,
-      "loss": 396.2762,
+      "loss": 353.8703,
       "step": 240
     },
     {
       "epoch": 13.333333333333334,
-      "grad_norm": 957.3853149414062,
+      "grad_norm": 873.0985107421875,
       "learning_rate": 5.580000000000001e-06,
-      "loss": 317.4443,
+      "loss": 262.4233,
       "step": 280
     },
     {
       "epoch": 15.238095238095237,
-      "grad_norm": 2215.865966796875,
+      "grad_norm": 1516.567626953125,
       "learning_rate": 6.380000000000001e-06,
-      "loss": 244.5671,
+      "loss": 205.628,
       "step": 320
     },
     {
       "epoch": 17.142857142857142,
-      "grad_norm": 494.6834716796875,
+      "grad_norm": 5139.3779296875,
       "learning_rate": 7.180000000000001e-06,
-      "loss": 200.2045,
+      "loss": 174.0845,
       "step": 360
     },
     {
       "epoch": 19.047619047619047,
-      "grad_norm": 389.99395751953125,
+      "grad_norm": 480.6575622558594,
       "learning_rate": 7.980000000000002e-06,
-      "loss": 143.4672,
+      "loss": 142.4243,
       "step": 400
     },
     {
       "epoch": 19.047619047619047,
-      "eval_loss": 212.24693298339844,
-      "eval_runtime": 7.6261,
-      "eval_samples_per_second": 15.211,
-      "eval_steps_per_second": 0.525,
+      "eval_loss": 214.48715209960938,
+      "eval_runtime": 7.9717,
+      "eval_samples_per_second": 14.551,
+      "eval_steps_per_second": 0.502,
       "step": 400
     },
     {
       "epoch": 20.952380952380953,
-      "grad_norm": 296.1668395996094,
+      "grad_norm": 366.5032958984375,
       "learning_rate": 8.78e-06,
-      "loss": 108.6625,
+      "loss": 115.4496,
       "step": 440
     },
     {
       "epoch": 22.857142857142858,
-      "grad_norm": 270.3917541503906,
+      "grad_norm": 435.5931701660156,
       "learning_rate": 9.58e-06,
-      "loss": 86.1836,
+      "loss": 92.9752,
       "step": 480
     },
     {
       "epoch": 24.761904761904763,
-      "grad_norm": 252.61672973632812,
+      "grad_norm": 232.87234497070312,
       "learning_rate": 9.993604266900276e-06,
-      "loss": 66.8246,
+      "loss": 78.7805,
       "step": 520
     },
     {
       "epoch": 26.666666666666668,
-      "grad_norm": 176.43890380859375,
+      "grad_norm": 188.52627563476562,
       "learning_rate": 9.938441702975689e-06,
-      "loss": 52.6602,
+      "loss": 59.0797,
       "step": 560
     },
     {
       "epoch": 28.571428571428573,
-      "grad_norm": 143.43080139160156,
+      "grad_norm": 177.4958038330078,
       "learning_rate": 9.827324621255206e-06,
-      "loss": 45.2324,
+      "loss": 46.9344,
       "step": 600
     },
     {
       "epoch": 28.571428571428573,
-      "eval_loss": 131.0333709716797,
-      "eval_runtime": 7.5366,
-      "eval_samples_per_second": 15.392,
-      "eval_steps_per_second": 0.531,
+      "eval_loss": 129.46548461914062,
+      "eval_runtime": 7.2365,
+      "eval_samples_per_second": 16.03,
+      "eval_steps_per_second": 0.553,
       "step": 600
     },
     {
       "epoch": 30.476190476190474,
-      "grad_norm": 125.09796142578125,
+      "grad_norm": 135.09072875976562,
       "learning_rate": 9.661512022674852e-06,
-      "loss": 36.3571,
+      "loss": 39.1039,
       "step": 640
     },
     {
       "epoch": 32.38095238095238,
-      "grad_norm": 130.8366241455078,
+      "grad_norm": 128.56063842773438,
       "learning_rate": 9.4428826302319e-06,
-      "loss": 30.4001,
+      "loss": 32.5044,
       "step": 680
     },
     {
       "epoch": 34.285714285714285,
-      "grad_norm": 98.4572525024414,
+      "grad_norm": 99.94125366210938,
       "learning_rate": 9.173913602302594e-06,
-      "loss": 27.7692,
+      "loss": 27.9798,
       "step": 720
     },
     {
       "epoch": 36.19047619047619,
-      "grad_norm": 118.89614868164062,
+      "grad_norm": 128.35171508789062,
       "learning_rate": 8.857652465447947e-06,
-      "loss": 24.9675,
+      "loss": 24.9475,
       "step": 760
     },
     {
       "epoch": 38.095238095238095,
-      "grad_norm": 89.57968139648438,
+      "grad_norm": 113.99468231201172,
       "learning_rate": 8.497682584720065e-06,
-      "loss": 24.3789,
+      "loss": 22.38,
       "step": 800
     },
     {
       "epoch": 38.095238095238095,
-      "eval_loss": 80.99005126953125,
-      "eval_runtime": 7.2774,
-      "eval_samples_per_second": 15.94,
-      "eval_steps_per_second": 0.55,
+      "eval_loss": 75.64146423339844,
+      "eval_runtime": 7.294,
+      "eval_samples_per_second": 15.904,
+      "eval_steps_per_second": 0.548,
       "step": 800
     },
     {
       "epoch": 40.0,
-      "grad_norm": 132.4896240234375,
+      "grad_norm": 190.8120574951172,
       "learning_rate": 8.098082562704253e-06,
-      "loss": 20.1034,
+      "loss": 19.868,
       "step": 840
     },
     {
       "epoch": 41.904761904761905,
-      "grad_norm": 81.84835815429688,
+      "grad_norm": 58.219482421875,
       "learning_rate": 7.663380027322022e-06,
-      "loss": 18.1971,
+      "loss": 22.0343,
       "step": 880
     },
     {
       "epoch": 43.80952380952381,
-      "grad_norm": 55.926551818847656,
+      "grad_norm": 89.1216049194336,
       "learning_rate": 7.1985003319975696e-06,
-      "loss": 17.9396,
+      "loss": 16.9501,
       "step": 920
     },
     {
       "epoch": 45.714285714285715,
-      "grad_norm": 64.777587890625,
+      "grad_norm": 147.2601776123047,
       "learning_rate": 6.708710749435427e-06,
-      "loss": 15.3426,
+      "loss": 15.9704,
       "step": 960
     },
     {
       "epoch": 47.61904761904762,
-      "grad_norm": 53.17697525024414,
+      "grad_norm": 103.56001281738281,
       "learning_rate": 6.199560791316009e-06,
-      "loss": 16.3762,
+      "loss": 15.38,
       "step": 1000
     },
     {
       "epoch": 47.61904761904762,
-      "eval_loss": 65.57749938964844,
-      "eval_runtime": 7.054,
-      "eval_samples_per_second": 16.445,
-      "eval_steps_per_second": 0.567,
+      "eval_loss": 79.28771209716797,
+      "eval_runtime": 7.3262,
+      "eval_samples_per_second": 15.834,
+      "eval_steps_per_second": 0.546,
       "step": 1000
     },
     {
       "epoch": 49.523809523809526,
-      "grad_norm": 140.15521240234375,
+      "grad_norm": 49.25249099731445,
       "learning_rate": 5.676819330110808e-06,
-      "loss": 13.8692,
+      "loss": 14.5406,
       "step": 1040
     },
     {
       "epoch": 51.42857142857143,
-      "grad_norm": 61.619049072265625,
+      "grad_norm": 116.98213195800781,
       "learning_rate": 5.14640923545217e-06,
-      "loss": 15.6367,
+      "loss": 14.0173,
       "step": 1080
     },
     {
       "epoch": 53.333333333333336,
-      "grad_norm": 147.7701873779297,
+      "grad_norm": 72.93340301513672,
       "learning_rate": 4.614340265653695e-06,
-      "loss": 13.0418,
+      "loss": 13.7052,
       "step": 1120
     },
     {
       "epoch": 55.23809523809524,
-      "grad_norm": 77.45187377929688,
+      "grad_norm": 105.87384796142578,
       "learning_rate": 4.086640974747176e-06,
-      "loss": 12.7083,
+      "loss": 15.4002,
       "step": 1160
     },
     {
       "epoch": 57.142857142857146,
-      "grad_norm": 215.86756896972656,
+      "grad_norm": 59.23909378051758,
       "learning_rate": 3.569290406556519e-06,
-      "loss": 13.65,
+      "loss": 14.1171,
       "step": 1200
     },
     {
       "epoch": 57.142857142857146,
-      "eval_loss": 58.122344970703125,
-      "eval_runtime": 7.0277,
-      "eval_samples_per_second": 16.506,
-      "eval_steps_per_second": 0.569,
+      "eval_loss": 64.76508331298828,
+      "eval_runtime": 7.1521,
+      "eval_samples_per_second": 16.219,
+      "eval_steps_per_second": 0.559,
       "step": 1200
     },
     {
       "epoch": 59.04761904761905,
-      "grad_norm": 76.79309844970703,
+      "grad_norm": 98.89187622070312,
       "learning_rate": 3.068150349742125e-06,
-      "loss": 11.9538,
+      "loss": 12.3597,
       "step": 1240
     },
     {
       "epoch": 60.95238095238095,
-      "grad_norm": 52.94865417480469,
+      "grad_norm": 179.5811004638672,
       "learning_rate": 2.5888989213931047e-06,
-      "loss": 11.5392,
+      "loss": 12.125,
       "step": 1280
     },
     {
       "epoch": 62.857142857142854,
-      "grad_norm": 153.39662170410156,
+      "grad_norm": 41.74519729614258,
       "learning_rate": 2.1369662316918095e-06,
-      "loss": 12.1913,
+      "loss": 11.7567,
       "step": 1320
     },
     {
       "epoch": 64.76190476190476,
-      "grad_norm": 139.7332305908203,
+      "grad_norm": 60.97040557861328,
       "learning_rate": 1.717472858595694e-06,
-      "loss": 11.2592,
+      "loss": 12.3442,
       "step": 1360
     },
     {
       "epoch": 66.66666666666667,
-      "grad_norm": 39.22611999511719,
+      "grad_norm": 197.75689697265625,
       "learning_rate": 1.335171829642984e-06,
-      "loss": 11.1986,
+      "loss": 11.2475,
       "step": 1400
     },
     {
       "epoch": 66.66666666666667,
-      "eval_loss": 60.4874382019043,
-      "eval_runtime": 7.0545,
-      "eval_samples_per_second": 16.443,
-      "eval_steps_per_second": 0.567,
+      "eval_loss": 67.15937805175781,
+      "eval_runtime": 7.2137,
+      "eval_samples_per_second": 16.08,
+      "eval_steps_per_second": 0.554,
       "step": 1400
     },
     {
       "epoch": 68.57142857142857,
-      "grad_norm": 80.5909423828125,
+      "grad_norm": 44.10291290283203,
       "learning_rate": 9.943947682514538e-07,
-      "loss": 11.4891,
+      "loss": 11.0752,
       "step": 1440
     },
     {
       "epoch": 70.47619047619048,
-      "grad_norm": 92.74817657470703,
+      "grad_norm": 54.24775314331055,
       "learning_rate": 6.990028146942752e-07,
-      "loss": 11.4975,
+      "loss": 10.9762,
       "step": 1480
     },
     {
       "epoch": 72.38095238095238,
-      "grad_norm": 58.092369079589844,
+      "grad_norm": 41.610679626464844,
       "learning_rate": 4.523428778378852e-07,
-      "loss": 11.2603,
+      "loss": 10.9561,
       "step": 1520
     },
     {
       "epoch": 74.28571428571429,
-      "grad_norm": 171.13815307617188,
+      "grad_norm": 109.53748321533203,
       "learning_rate": 2.5720971332717783e-07,
-      "loss": 10.5585,
+      "loss": 10.7585,
       "step": 1560
     },
     {
       "epoch": 76.19047619047619,
-      "grad_norm": 49.394981384277344,
+      "grad_norm": 49.795623779296875,
       "learning_rate": 1.158142578873156e-07,
-      "loss": 10.9979,
+      "loss": 11.5847,
       "step": 1600
     },
     {
       "epoch": 76.19047619047619,
-      "eval_loss": 56.9376335144043,
-      "eval_runtime": 7.1981,
-      "eval_samples_per_second": 16.115,
-      "eval_steps_per_second": 0.556,
+      "eval_loss": 62.760154724121094,
+      "eval_runtime": 7.2145,
+      "eval_samples_per_second": 16.079,
+      "eval_steps_per_second": 0.554,
       "step": 1600
     },
     {
       "epoch": 78.0952380952381,
-      "grad_norm": 55.77260971069336,
+      "grad_norm": 58.66199493408203,
       "learning_rate": 2.9758578527189776e-08,
-      "loss": 10.6299,
+      "loss": 11.5135,
       "step": 1640
     },
     {
       "epoch": 80.0,
-      "grad_norm": 59.74852752685547,
+      "grad_norm": 53.418094635009766,
       "learning_rate": 1.7720480054994606e-11,
-      "loss": 10.527,
+      "loss": 10.746,
       "step": 1680
     }
   ],

checkpoint-1680/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c0e178d29d8bd1005e0ebacb559c58b57ef56dd83b9d009bde5540a5128a1d4a
+oid sha256:ba4185c232729e9f26479512d5ae60057198316f129ed28036d583f86d686e96
 size 5777

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8fc83e9ffb9854d6f50d18b71d3610cb9b411e48e241f6e96354f2fd615e20c4
+oid sha256:26f7b879ae25a9c83fa3e849fdb085366c9c812655437eeedbdf446b895a6cc5
 size 306699044

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c0e178d29d8bd1005e0ebacb559c58b57ef56dd83b9d009bde5540a5128a1d4a
+oid sha256:ba4185c232729e9f26479512d5ae60057198316f129ed28036d583f86d686e96
 size 5777