Upload dpo_metrics.jsonl with huggingface_hub
Browse files- dpo_metrics.jsonl +63 -0
dpo_metrics.jsonl
CHANGED
|
@@ -290,3 +290,66 @@
|
|
| 290 |
{"timestamp": 1787716190.460649, "elapsed_hours": 0.3339935287502077, "event": "dpo_step", "step": 2760, "loss": 0.709025399684906, "pref_accuracy": 0.48, "chosen_reward": -0.1113433837890625, "rejected_reward": 0.1878509521484375, "lr": 1.531707128387244e-06, "tokens_per_sec": 151633.15775519225, "vram_gb": 4.5285258293151855}
|
| 291 |
{"timestamp": 1787716195.1347816, "elapsed_hours": 0.3352918988466263, "event": "dpo_step", "step": 2770, "loss": 0.723406343460083, "pref_accuracy": 0.47, "chosen_reward": 0.11997070908546448, "rejected_reward": 0.18498899042606354, "lr": 1.5292084505792967e-06, "tokens_per_sec": 151588.56912585793, "vram_gb": 4.592715740203857}
|
| 292 |
{"timestamp": 1787716199.3872063, "elapsed_hours": 0.33647312813334995, "event": "dpo_step", "step": 2780, "loss": 0.7145835375785827, "pref_accuracy": 0.52, "chosen_reward": -0.01452636905014515, "rejected_reward": -0.5156311392784119, "lr": 1.5267054832052494e-06, "tokens_per_sec": 151597.50878709852, "vram_gb": 4.956247329711914}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 290 |
{"timestamp": 1787716190.460649, "elapsed_hours": 0.3339935287502077, "event": "dpo_step", "step": 2760, "loss": 0.709025399684906, "pref_accuracy": 0.48, "chosen_reward": -0.1113433837890625, "rejected_reward": 0.1878509521484375, "lr": 1.531707128387244e-06, "tokens_per_sec": 151633.15775519225, "vram_gb": 4.5285258293151855}
|
| 291 |
{"timestamp": 1787716195.1347816, "elapsed_hours": 0.3352918988466263, "event": "dpo_step", "step": 2770, "loss": 0.723406343460083, "pref_accuracy": 0.47, "chosen_reward": 0.11997070908546448, "rejected_reward": 0.18498899042606354, "lr": 1.5292084505792967e-06, "tokens_per_sec": 151588.56912585793, "vram_gb": 4.592715740203857}
|
| 292 |
{"timestamp": 1787716199.3872063, "elapsed_hours": 0.33647312813334995, "event": "dpo_step", "step": 2780, "loss": 0.7145835375785827, "pref_accuracy": 0.52, "chosen_reward": -0.01452636905014515, "rejected_reward": -0.5156311392784119, "lr": 1.5267054832052494e-06, "tokens_per_sec": 151597.50878709852, "vram_gb": 4.956247329711914}
|
| 293 |
+
{"timestamp": 1787716203.7447147, "elapsed_hours": 0.337683547006713, "event": "dpo_step", "step": 2790, "loss": 0.6948308968544006, "pref_accuracy": 0.59, "chosen_reward": 0.117498017847538, "rejected_reward": 0.007831955328583717, "lr": 1.5241982512800895e-06, "tokens_per_sec": 151593.17409687955, "vram_gb": 3.953242301940918}
|
| 294 |
+
{"timestamp": 1787716207.3275533, "elapsed_hours": 0.3386787798669603, "event": "dpo_step", "step": 2800, "loss": 0.691988525390625, "pref_accuracy": 0.6, "chosen_reward": -0.16420288383960724, "rejected_reward": 0.03130493313074112, "lr": 1.5191710940492346e-06, "tokens_per_sec": 151685.94334470638, "vram_gb": 4.5285258293151855}
|
| 295 |
+
{"timestamp": 1787716210.4555736, "elapsed_hours": 0.33954767439100475, "event": "sample", "step": 2800, "sample_text": "```python\ndef fibonacci(n):\n if n == 0:\n return 1\n elif n == 1:\n return n - 1\n else:\n return fibonacci(n-1) + fibonacci(n-2)\n\n# Example usage:\nprint(fibonacci(10)) # Output: 1\nprint(fibonacci(20)) # Output: 2\nprint(fibonacci(30)) # Output: 3\nprint(fibonacci(40)) # Output: 2"}
|
| 296 |
+
{"timestamp": 1787716214.1178653, "elapsed_hours": 0.3405649777253469, "event": "dpo_step", "step": 2810, "loss": 0.6770584976673126, "pref_accuracy": 0.63, "chosen_reward": 0.04926758259534836, "rejected_reward": 0.19403076171875, "lr": 1.5166512189856146e-06, "tokens_per_sec": 151377.99140271216, "vram_gb": 4.9414825439453125}
|
| 297 |
+
{"timestamp": 1787716217.6866047, "elapsed_hours": 0.3415562942292955, "event": "dpo_step", "step": 2820, "loss": 0.6810489654541015, "pref_accuracy": 0.59, "chosen_reward": 0.05771942064166069, "rejected_reward": -0.05204048007726669, "lr": 1.5141271798545303e-06, "tokens_per_sec": 151472.3536719704, "vram_gb": 4.300653457641602}
|
| 298 |
+
{"timestamp": 1787716221.0681295, "elapsed_hours": 0.34249560667408835, "event": "dpo_step", "step": 2830, "loss": 0.696682162284851, "pref_accuracy": 0.51, "chosen_reward": 0.037613678723573685, "rejected_reward": 0.0814208984375, "lr": 1.5115990018815626e-06, "tokens_per_sec": 151589.35509224425, "vram_gb": 4.956491470336914}
|
| 299 |
+
{"timestamp": 1787716224.931451, "elapsed_hours": 0.34356875154707167, "event": "dpo_step", "step": 2840, "loss": 0.7072904360294342, "pref_accuracy": 0.42, "chosen_reward": 0.338217169046402, "rejected_reward": -0.0837913528084755, "lr": 1.5065303305188733e-06, "tokens_per_sec": 151646.16437191752, "vram_gb": 4.5285258293151855}
|
| 300 |
+
{"timestamp": 1787716229.3695722, "elapsed_hours": 0.3448015628920661, "event": "dpo_step", "step": 2850, "loss": 0.710717808008194, "pref_accuracy": 0.4, "chosen_reward": -0.085968017578125, "rejected_reward": 0.1399078369140625, "lr": 1.503989887786124e-06, "tokens_per_sec": 151631.82204648934, "vram_gb": 4.062510967254639}
|
| 301 |
+
{"timestamp": 1787716233.4298744, "elapsed_hours": 0.34592942469649846, "event": "dpo_step", "step": 2860, "loss": 0.7240729796886444, "pref_accuracy": 0.34, "chosen_reward": 0.04141846299171448, "rejected_reward": 0.084716796875, "lr": 1.5014454075249313e-06, "tokens_per_sec": 151663.9389374586, "vram_gb": 4.94575834274292}
|
| 302 |
+
{"timestamp": 1787716237.758669, "elapsed_hours": 0.3471318676074346, "event": "dpo_step", "step": 2870, "loss": 0.7224322533607483, "pref_accuracy": 0.36, "chosen_reward": -0.12650223076343536, "rejected_reward": 0.05508880689740181, "lr": 1.4988969151651681e-06, "tokens_per_sec": 151663.00072921682, "vram_gb": 4.173646450042725}
|
| 303 |
+
{"timestamp": 1787716242.382949, "elapsed_hours": 0.3484163898229599, "event": "dpo_step", "step": 2880, "loss": 0.719368988275528, "pref_accuracy": 0.4, "chosen_reward": -0.0550537109375, "rejected_reward": -0.0106201171875, "lr": 1.493787996069653e-06, "tokens_per_sec": 151626.0672316674, "vram_gb": 4.5285258293151855}
|
| 304 |
+
{"timestamp": 1787716246.3686023, "elapsed_hours": 0.3495235158337487, "event": "dpo_step", "step": 2890, "loss": 0.7187606316804885, "pref_accuracy": 0.45, "chosen_reward": -0.03991394117474556, "rejected_reward": -0.21451720595359802, "lr": 1.4912276203931159e-06, "tokens_per_sec": 151666.9340387943, "vram_gb": 4.947980880737305}
|
| 305 |
+
{"timestamp": 1787716250.6798093, "elapsed_hours": 0.3507210733493169, "event": "dpo_step", "step": 2900, "loss": 0.7133746689558029, "pref_accuracy": 0.51, "chosen_reward": -0.2033620923757553, "rejected_reward": -0.08902053534984589, "lr": 1.4886633347359255e-06, "tokens_per_sec": 151668.12729680983, "vram_gb": 4.956308364868164}
|
| 306 |
+
{"timestamp": 1787716254.5706215, "elapsed_hours": 0.3518018544382519, "event": "dpo_step", "step": 2910, "loss": 0.715048480629921, "pref_accuracy": 0.54, "chosen_reward": -0.03435669094324112, "rejected_reward": -0.013550949282944202, "lr": 1.4860951647258927e-06, "tokens_per_sec": 151720.03506050713, "vram_gb": 4.949965000152588}
|
| 307 |
+
{"timestamp": 1787716258.6684465, "elapsed_hours": 0.3529401391744614, "event": "dpo_step", "step": 2920, "loss": 0.7124473232030869, "pref_accuracy": 0.53, "chosen_reward": -0.019339753314852715, "rejected_reward": 0.0257644671946764, "lr": 1.4809472743523899e-06, "tokens_per_sec": 151746.70648986037, "vram_gb": 3.606423854827881}
|
| 308 |
+
{"timestamp": 1787716262.6589172, "elapsed_hours": 0.35404860330952537, "event": "dpo_step", "step": 2930, "loss": 0.7018657058477402, "pref_accuracy": 0.56, "chosen_reward": 0.03853759169578552, "rejected_reward": -0.19973145425319672, "lr": 1.4783676054376177e-06, "tokens_per_sec": 151786.08891806993, "vram_gb": 4.491017818450928}
|
| 309 |
+
{"timestamp": 1787716266.782756, "elapsed_hours": 0.3551941141155031, "event": "dpo_step", "step": 2940, "loss": 0.7034299921989441, "pref_accuracy": 0.53, "chosen_reward": -0.09153442829847336, "rejected_reward": -0.0523376502096653, "lr": 1.4757841550668855e-06, "tokens_per_sec": 151809.27035567776, "vram_gb": 4.956247329711914}
|
| 310 |
+
{"timestamp": 1787716270.5915108, "elapsed_hours": 0.35625210139486524, "event": "dpo_step", "step": 2950, "loss": 0.7058947098255157, "pref_accuracy": 0.48, "chosen_reward": -0.05718841403722763, "rejected_reward": 0.15091857314109802, "lr": 1.4731969490595381e-06, "tokens_per_sec": 151869.87275071422, "vram_gb": 3.965236186981201}
|
| 311 |
+
{"timestamp": 1787716274.6559398, "elapsed_hours": 0.357381109462844, "event": "dpo_step", "step": 2960, "loss": 0.6993654334545135, "pref_accuracy": 0.51, "chosen_reward": 0.07296905666589737, "rejected_reward": -0.0508270263671875, "lr": 1.4680113735997927e-06, "tokens_per_sec": 151899.70206291604, "vram_gb": 4.049484729766846}
|
| 312 |
+
{"timestamp": 1787716278.154623, "elapsed_hours": 0.35835296591122945, "event": "dpo_step", "step": 2970, "loss": 0.6888049232959748, "pref_accuracy": 0.56, "chosen_reward": 0.03417396545410156, "rejected_reward": -0.08084382861852646, "lr": 1.4654130559727225e-06, "tokens_per_sec": 151996.46605802554, "vram_gb": 4.0814361572265625}
|
| 313 |
+
{"timestamp": 1787716283.2827551, "elapsed_hours": 0.35977744711769954, "event": "dpo_step", "step": 2980, "loss": 0.7070292413234711, "pref_accuracy": 0.52, "chosen_reward": -0.26295167207717896, "rejected_reward": -0.20856933295726776, "lr": 1.4628110863591752e-06, "tokens_per_sec": 151899.95442526066, "vram_gb": 4.946849822998047}
|
| 314 |
+
{"timestamp": 1787716286.9418473, "elapsed_hours": 0.36079386168056066, "event": "dpo_step", "step": 2990, "loss": 0.7143213951587677, "pref_accuracy": 0.52, "chosen_reward": -0.14971008896827698, "rejected_reward": 0.25667116045951843, "lr": 1.4602054907635795e-06, "tokens_per_sec": 151977.1525937744, "vram_gb": 4.943683624267578}
|
| 315 |
+
{"timestamp": 1787716290.2488468, "elapsed_hours": 0.36171247257126704, "event": "dpo_step", "step": 3000, "loss": 0.7128161835670471, "pref_accuracy": 0.52, "chosen_reward": -0.05940704420208931, "rejected_reward": 0.06441688537597656, "lr": 1.4549835258248925e-06, "tokens_per_sec": 152095.33339772362, "vram_gb": 4.041985988616943}
|
| 316 |
+
{"timestamp": 1787716293.3627708, "elapsed_hours": 0.362577451467514, "event": "sample", "step": 3000, "sample_text": "Here is a Python function that returns the nth Fibonacci number:\n\n```python\ndef fibonacci(n):\n if n == 1:\n return 1\n elif n == 2:\n return 2\n else:\n return fibonacci(n-1) + fibonacci(n-2)\n```\n\nThis function first checks if the input `n` is a valid integer. If it is, it returns 1. Otherwise, it returns 2. If it is not a valid integer, it returns 2.\n\nThe function then"}
|
| 317 |
+
{"timestamp": 1787716304.1520226, "elapsed_hours": 0.3655744658576118, "event": "dpo_step", "step": 3010, "loss": 0.7045691466331482, "pref_accuracy": 0.53, "chosen_reward": -0.2557128965854645, "rejected_reward": 0.00301513634622097, "lr": 1.4523672086708098e-06, "tokens_per_sec": 150978.47597757008, "vram_gb": 4.956186294555664}
|
| 318 |
+
{"timestamp": 1787716308.6998017, "elapsed_hours": 0.36683773782518175, "event": "dpo_step", "step": 3020, "loss": 0.7093925547599792, "pref_accuracy": 0.51, "chosen_reward": 0.07018127292394638, "rejected_reward": 0.12701110541820526, "lr": 1.4497473699121762e-06, "tokens_per_sec": 150954.6392869168, "vram_gb": 4.62000846862793}
|
| 319 |
+
{"timestamp": 1787716313.2258596, "elapsed_hours": 0.3680949762132433, "event": "dpo_step", "step": 3030, "loss": 0.702077043056488, "pref_accuracy": 0.49, "chosen_reward": 0.07370606064796448, "rejected_reward": -0.04085388779640198, "lr": 1.4471240357320066e-06, "tokens_per_sec": 150933.45527928346, "vram_gb": 4.955257415771484}
|
| 320 |
+
{"timestamp": 1787716315.8524861, "elapsed_hours": 0.36882459468311735, "event": "dpo_step", "step": 3040, "loss": 0.7007862675189972, "pref_accuracy": 0.46, "chosen_reward": -0.12147674709558487, "rejected_reward": -0.1478775143623352, "lr": 1.4418669860135286e-06, "tokens_per_sec": 151129.85930253513, "vram_gb": 3.7376184463500977}
|
| 321 |
+
{"timestamp": 1787716319.734019, "elapsed_hours": 0.3699027982685301, "event": "dpo_step", "step": 3050, "loss": 0.7028797268867493, "pref_accuracy": 0.47, "chosen_reward": 0.32304078340530396, "rejected_reward": -0.07535629719495773, "lr": 1.4392333230148718e-06, "tokens_per_sec": 151181.85468449947, "vram_gb": 4.956308364868164}
|
| 322 |
+
{"timestamp": 1787716324.148204, "elapsed_hours": 0.37112896071539986, "event": "dpo_step", "step": 3060, "loss": 0.712312968969345, "pref_accuracy": 0.45, "chosen_reward": -0.011584475636482239, "rejected_reward": -0.03542175143957138, "lr": 1.4365962696734558e-06, "tokens_per_sec": 151172.82086399346, "vram_gb": 4.956308364868164}
|
| 323 |
+
{"timestamp": 1787716328.7201698, "elapsed_hours": 0.3723989513185289, "event": "dpo_step", "step": 3070, "loss": 0.7117994022369385, "pref_accuracy": 0.48, "chosen_reward": -0.15458986163139343, "rejected_reward": -0.16804200410842896, "lr": 1.4339558523443419e-06, "tokens_per_sec": 151145.92839096364, "vram_gb": 4.956562042236328}
|
| 324 |
+
{"timestamp": 1787716332.2288978, "elapsed_hours": 0.3733735978603363, "event": "dpo_step", "step": 3080, "loss": 0.7074272060394287, "pref_accuracy": 0.49, "chosen_reward": -0.02706909365952015, "rejected_reward": 0.05615844950079918, "lr": 1.4286650313110975e-06, "tokens_per_sec": 151239.60766659657, "vram_gb": 3.6484060287475586}
|
| 325 |
+
{"timestamp": 1787716335.287766, "elapsed_hours": 0.37422328352928164, "event": "dpo_step", "step": 3090, "loss": 0.6954503905773163, "pref_accuracy": 0.56, "chosen_reward": 0.06009521335363388, "rejected_reward": 0.146759033203125, "lr": 1.4260146804841332e-06, "tokens_per_sec": 151383.69253787323, "vram_gb": 3.999722480773926}
|
| 326 |
+
{"timestamp": 1787716340.3583295, "elapsed_hours": 0.3756317734056049, "event": "dpo_step", "step": 3100, "loss": 0.6937197017669677, "pref_accuracy": 0.55, "chosen_reward": 0.3156494200229645, "rejected_reward": 0.01690063625574112, "lr": 1.4233610714232746e-06, "tokens_per_sec": 151300.10007631948, "vram_gb": 4.956308364868164}
|
| 327 |
+
{"timestamp": 1787716344.8908663, "elapsed_hours": 0.37689081132411956, "event": "dpo_step", "step": 3110, "loss": 0.6902731275558471, "pref_accuracy": 0.53, "chosen_reward": -0.11914673447608948, "rejected_reward": -0.08466797322034836, "lr": 1.4207042306490419e-06, "tokens_per_sec": 151277.5268030101, "vram_gb": 4.716411590576172}
|
| 328 |
+
{"timestamp": 1787716348.8275185, "elapsed_hours": 0.37798432581954533, "event": "dpo_step", "step": 3120, "loss": 0.7093745458126068, "pref_accuracy": 0.46, "chosen_reward": 0.07206955552101135, "rejected_reward": -0.007023239508271217, "lr": 1.4153809602037612e-06, "tokens_per_sec": 151321.8041714544, "vram_gb": 3.4362473487854004}
|
| 329 |
+
{"timestamp": 1787716352.2263265, "elapsed_hours": 0.3789284391535653, "event": "dpo_step", "step": 3130, "loss": 0.7067919623851776, "pref_accuracy": 0.5, "chosen_reward": -0.05022583156824112, "rejected_reward": -0.01428527757525444, "lr": 1.4127145837341835e-06, "tokens_per_sec": 151425.92662074638, "vram_gb": 3.784456729888916}
|
| 330 |
+
{"timestamp": 1787716357.1072357, "elapsed_hours": 0.3802842473321491, "event": "dpo_step", "step": 3140, "loss": 0.7144850903749466, "pref_accuracy": 0.46, "chosen_reward": -0.2713455259799957, "rejected_reward": -0.20197682082653046, "lr": 1.4100450819536392e-06, "tokens_per_sec": 151364.34166666426, "vram_gb": 4.956613540649414}
|
| 331 |
+
{"timestamp": 1787716361.460268, "elapsed_hours": 0.3814934229188495, "event": "dpo_step", "step": 3150, "loss": 0.7070700079202652, "pref_accuracy": 0.48, "chosen_reward": -0.045074462890625, "rejected_reward": -0.3749023675918579, "lr": 1.4073724815414835e-06, "tokens_per_sec": 151361.7423278586, "vram_gb": 4.955942153930664}
|
| 332 |
+
{"timestamp": 1787716365.3280973, "elapsed_hours": 0.3825678199529648, "event": "dpo_step", "step": 3160, "loss": 0.7257068186998368, "pref_accuracy": 0.44, "chosen_reward": -0.34047242999076843, "rejected_reward": -0.07442321628332138, "lr": 1.4020180916943281e-06, "tokens_per_sec": 151412.86348534806, "vram_gb": 4.5285258293151855}
|
| 333 |
+
{"timestamp": 1787716369.268627, "elapsed_hours": 0.38366241157054903, "event": "dpo_step", "step": 3170, "loss": 0.718284804224968, "pref_accuracy": 0.45, "chosen_reward": 0.21360473334789276, "rejected_reward": 0.23760223388671875, "lr": 1.3993363557718096e-06, "tokens_per_sec": 151455.6722665475, "vram_gb": 4.946487903594971}
|
| 334 |
+
{"timestamp": 1787716374.1378982, "elapsed_hours": 0.38501498692565495, "event": "dpo_step", "step": 3180, "loss": 0.7204569643735885, "pref_accuracy": 0.44, "chosen_reward": 0.029933547601103783, "rejected_reward": -0.027486611157655716, "lr": 1.396651628242107e-06, "tokens_per_sec": 151396.01340629105, "vram_gb": 4.500504970550537}
|
| 335 |
+
{"timestamp": 1787716377.542994, "elapsed_hours": 0.385960846874449, "event": "dpo_step", "step": 3190, "loss": 0.7271189272403717, "pref_accuracy": 0.43, "chosen_reward": -0.11035767197608948, "rejected_reward": 0.16318587958812714, "lr": 1.3939639359367431e-06, "tokens_per_sec": 151497.36123303475, "vram_gb": 4.128221035003662}
|
| 336 |
+
{"timestamp": 1787716380.8286288, "elapsed_hours": 0.3868735231955846, "event": "dpo_step", "step": 3200, "loss": 0.7266518890857696, "pref_accuracy": 0.46, "chosen_reward": 0.03997344896197319, "rejected_reward": 0.12385411560535431, "lr": 1.3885797644730042e-06, "tokens_per_sec": 151611.28747390865, "vram_gb": 4.5285258293151855}
|
| 337 |
+
{"timestamp": 1787716383.921151, "elapsed_hours": 0.3877325571245617, "event": "sample", "step": 3200, "sample_text": "Here is a Python function that returns the nth Fibonacci number:\n\n```python\ndef fibonacci(n):\n if n < 0:\n return 0\n else:\n return fibonacci(n-1) + fibonacci(n-2)\n```\n\nThis function first checks if n is equal to 0. If it is, then it returns 0. Otherwise, it returns 1. If n is not equal to 0, it returns 1.\n\nThis function has a time complexity of O(n^2), which means it has a time complexity of O(n), where"}
|
| 338 |
+
{"timestamp": 1787716387.2219503, "elapsed_hours": 0.38864944577217103, "event": "dpo_step", "step": 3210, "loss": 0.7206989371776581, "pref_accuracy": 0.48, "chosen_reward": 0.032018281519412994, "rejected_reward": -0.08554687350988388, "lr": 1.385883339124752e-06, "tokens_per_sec": 151385.36479610438, "vram_gb": 3.7072176933288574}
|
| 339 |
+
{"timestamp": 1787716389.3341217, "elapsed_hours": 0.38923615998691985, "event": "dpo_step", "step": 3220, "loss": 0.7153461074829102, "pref_accuracy": 0.5, "chosen_reward": 0.041828155517578125, "rejected_reward": 0.045894622802734375, "lr": 1.3831840566205456e-06, "tokens_per_sec": 151626.50558005276, "vram_gb": 4.10936975479126}
|
| 340 |
+
{"timestamp": 1787716394.3804169, "elapsed_hours": 0.3906379086441464, "event": "dpo_step", "step": 3230, "loss": 0.719692451953888, "pref_accuracy": 0.47, "chosen_reward": 0.03710480034351349, "rejected_reward": 0.07789306342601776, "lr": 1.3804819439373724e-06, "tokens_per_sec": 151547.90389428643, "vram_gb": 3.8205485343933105}
|
| 341 |
+
{"timestamp": 1787716399.0698974, "elapsed_hours": 0.391940542128351, "event": "dpo_step", "step": 3240, "loss": 0.7052076476812362, "pref_accuracy": 0.49, "chosen_reward": -0.07687301933765411, "rejected_reward": 0.0133209228515625, "lr": 1.3750693360832283e-06, "tokens_per_sec": 151508.4295879709, "vram_gb": 4.5285258293151855}
|
| 342 |
+
{"timestamp": 1787716402.3623235, "elapsed_hours": 0.39285510500272114, "event": "dpo_step", "step": 3250, "loss": 0.7132323342561722, "pref_accuracy": 0.45, "chosen_reward": -0.05066680535674095, "rejected_reward": 0.01431579515337944, "lr": 1.372358895006577e-06, "tokens_per_sec": 151619.85810130296, "vram_gb": 3.751481056213379}
|
| 343 |
+
{"timestamp": 1787716406.779552, "elapsed_hours": 0.3940821129745907, "event": "dpo_step", "step": 3260, "loss": 0.7097245907783508, "pref_accuracy": 0.46, "chosen_reward": 0.313436895608902, "rejected_reward": -0.4630935788154602, "lr": 1.3696457319390575e-06, "tokens_per_sec": 151609.65669493197, "vram_gb": 4.956186294555664}
|
| 344 |
+
{"timestamp": 1787716410.2545326, "elapsed_hours": 0.3950473853614595, "event": "dpo_step", "step": 3270, "loss": 0.7071872651576996, "pref_accuracy": 0.46, "chosen_reward": -0.011499021202325821, "rejected_reward": -0.251596063375473, "lr": 1.3669298739963814e-06, "tokens_per_sec": 151700.63856735307, "vram_gb": 4.953144550323486}
|
| 345 |
+
{"timestamp": 1787716414.1891162, "elapsed_hours": 0.39614032520188225, "event": "dpo_step", "step": 3280, "loss": 0.6986054742336273, "pref_accuracy": 0.51, "chosen_reward": -0.01387634128332138, "rejected_reward": 0.2991470396518707, "lr": 1.3614901820827945e-06, "tokens_per_sec": 151741.91415348384, "vram_gb": 4.5285258293151855}
|
| 346 |
+
{"timestamp": 1787716419.3869863, "elapsed_hours": 0.3975841779841317, "event": "dpo_step", "step": 3290, "loss": 0.6967894321680069, "pref_accuracy": 0.57, "chosen_reward": 0.013580322265625, "rejected_reward": -0.043517302721738815, "lr": 1.3587664024768856e-06, "tokens_per_sec": 151648.1053747315, "vram_gb": 4.956308364868164}
|
| 347 |
+
{"timestamp": 1787716424.1546903, "elapsed_hours": 0.39890854014290705, "event": "dpo_step", "step": 3300, "loss": 0.6991162914037704, "pref_accuracy": 0.56, "chosen_reward": 0.03875732421875, "rejected_reward": -0.00820312462747097, "lr": 1.3560400367252787e-06, "tokens_per_sec": 151600.6822816672, "vram_gb": 4.956125259399414}
|
| 348 |
+
{"timestamp": 1787716427.6749463, "elapsed_hours": 0.3998863890435961, "event": "dpo_step", "step": 3310, "loss": 0.7019866073131561, "pref_accuracy": 0.54, "chosen_reward": -0.15940704941749573, "rejected_reward": 0.001773834228515625, "lr": 1.3533111120756349e-06, "tokens_per_sec": 151685.77223107536, "vram_gb": 3.8600525856018066}
|
| 349 |
+
{"timestamp": 1787716430.5013573, "elapsed_hours": 0.40067150327894424, "event": "dpo_step", "step": 3320, "loss": 0.6961434352397918, "pref_accuracy": 0.57, "chosen_reward": 0.26915284991264343, "rejected_reward": 0.21270447969436646, "lr": 1.347845695200475e-06, "tokens_per_sec": 151843.93603986446, "vram_gb": 4.5285258293151855}
|
| 350 |
+
{"timestamp": 1787716434.3300622, "elapsed_hours": 0.40173503239949543, "event": "dpo_step", "step": 3330, "loss": 0.6995329523086548, "pref_accuracy": 0.54, "chosen_reward": 0.10661621391773224, "rejected_reward": 0.1586044430732727, "lr": 1.3451092575970583e-06, "tokens_per_sec": 151895.43767862156, "vram_gb": 4.942780017852783}
|
| 351 |
+
{"timestamp": 1787716438.8104868, "elapsed_hours": 0.40297959480020734, "event": "dpo_step", "step": 3340, "loss": 0.711788227558136, "pref_accuracy": 0.49, "chosen_reward": -0.1344398558139801, "rejected_reward": -0.3830429017543793, "lr": 1.3423703703392587e-06, "tokens_per_sec": 151877.954224949, "vram_gb": 4.5221099853515625}
|
| 352 |
+
{"timestamp": 1787716443.8663483, "elapsed_hours": 0.40438400083118015, "event": "dpo_step", "step": 3350, "loss": 0.7094157755374908, "pref_accuracy": 0.49, "chosen_reward": 0.01430664025247097, "rejected_reward": 0.1682586669921875, "lr": 1.3396290607998786e-06, "tokens_per_sec": 151800.1612990872, "vram_gb": 4.956125259399414}
|
| 353 |
+
{"timestamp": 1787716448.2911553, "elapsed_hours": 0.40561311390664845, "event": "dpo_step", "step": 3360, "loss": 0.6989813315868377, "pref_accuracy": 0.51, "chosen_reward": 0.05009765923023224, "rejected_reward": 0.00017318129539489746, "lr": 1.3341392844883522e-06, "tokens_per_sec": 151788.9051220956, "vram_gb": 3.9625210762023926}
|
| 354 |
+
{"timestamp": 1787716451.2956932, "elapsed_hours": 0.4064477076795366, "event": "dpo_step", "step": 3370, "loss": 0.7055905616283417, "pref_accuracy": 0.49, "chosen_reward": 0.09941559284925461, "rejected_reward": 0.059179116040468216, "lr": 1.331390872581758e-06, "tokens_per_sec": 151926.00655195306, "vram_gb": 4.551896572113037}
|
| 355 |
+
{"timestamp": 1787716795.5775218, "elapsed_hours": 0.003705000281333923, "event": "dpo_start", "config": {"vocab_size": 50304, "d_model": 896, "n_layers": 18, "n_heads": 14, "n_kv_heads": 2, "max_len": 2048, "rope_theta": 500000.0, "window_size": 512, "pattern_mult": 2.5, "reason_mult": 0.75, "reason_depth": 2, "gate_hidden": 64, "gate_init": 0.0, "diff_lambda_init": 0.8, "tie_embeddings": true, "beta": 0.3, "dpo_lr": 5e-06, "dpo_min_lr": 2e-07, "wd": 0.0, "grad_clip": 1.0, "warmup": 50, "total_steps": 8000, "micro_batch": 2, "grad_accum": 8, "log_every": 10, "sample_every": 200, "ckpt_every": 500, "reason_prob": 0.0, "stop_corrupt_prob": 0.35}, "model_params": 213061285, "n_examples": 26000, "resumed": true, "resumed_from_step": 3000}
|