Training in progress, step 24, checkpoint

last-checkpoint/adapter_config.json

@@ -10,22 +10,22 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 16,
-  "lora_dropout": 0.1,
+  "lora_alpha": 32,
+  "lora_dropout": 0.05,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": null,
   "peft_type": "LORA",
-  "r": 8,
+  "r": 16,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "k_proj",
+    "gate_proj",
     "up_proj",
-    "v_proj",
     "down_proj",
     "q_proj",
-    "gate_proj",
+    "v_proj",
+    "k_proj",
     "o_proj"
   ],
   "task_type": "CAUSAL_LM",

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:65278dd1c880975399af59ced022b36ec8b756e4bef76d0024be8fc9e78457f4
-size 83945296
+oid sha256:f153213500f7967f5c5e9e61ed603bfa80255b7508db241aa248a2a5a8bca590
+size 167832240

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a3a0e23486cb88a362571ea7128834a5cea117d12a2d6acf6b88ef2f38c020c4
-size 168149074
+oid sha256:2830497fadb8a4e19b5c7af4fdc029a1b0a294506d5d50c365776cb488290401
+size 335922386

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c3663c1f4fd9998792e9716d39f30557c765edb480fe3e9a654c864730c56331
+oid sha256:01e2c00897b8e745718f5156c37cdc3f3ea9cd1e977ad43db263ac01587adda9
 size 14244

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7f8e3d30c4f47a519421ea0a44735662bdb9869afd75e3f7738f31c9ef193e41
+oid sha256:470670603e8fdc5330cdb9a9152c4fd9c3d8c5a74dd26bffbbb0d869d097eafa
 size 1064

last-checkpoint/trainer_state.json

@@ -1,16 +1,16 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.0016320239363510666,
-  "eval_steps": 3,
-  "global_step": 18,
+  "epoch": 0.0021760319151347554,
+  "eval_steps": 4,
+  "global_step": 24,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 9.066799646394814e-05,
-      "grad_norm": 0.9196107983589172,
+      "grad_norm": 1.305655837059021,
       "learning_rate": 2e-05,
       "loss": 0.9466,
       "step": 1
@@ -18,184 +18,226 @@
     {
       "epoch": 9.066799646394814e-05,
       "eval_loss": 1.1514371633529663,
-      "eval_runtime": 280.9406,
-      "eval_samples_per_second": 8.265,
-      "eval_steps_per_second": 8.265,
+      "eval_runtime": 281.757,
+      "eval_samples_per_second": 8.241,
+      "eval_steps_per_second": 8.241,
       "step": 1
     },
     {
       "epoch": 0.0001813359929278963,
-      "grad_norm": 1.2303106784820557,
+      "grad_norm": 1.7540494203567505,
       "learning_rate": 4e-05,
       "loss": 1.2689,
       "step": 2
     },
     {
       "epoch": 0.00027200398939184443,
-      "grad_norm": 0.953157901763916,
+      "grad_norm": 1.3806679248809814,
       "learning_rate": 6e-05,
-      "loss": 0.9876,
+      "loss": 0.9866,
       "step": 3
     },
     {
-      "epoch": 0.00027200398939184443,
-      "eval_loss": 1.1419087648391724,
-      "eval_runtime": 281.2312,
-      "eval_samples_per_second": 8.257,
-      "eval_steps_per_second": 8.257,
-      "step": 3
+      "epoch": 0.0003626719858557926,
+      "grad_norm": 1.6540546417236328,
+      "learning_rate": 8e-05,
+      "loss": 0.886,
+      "step": 4
     },
     {
       "epoch": 0.0003626719858557926,
-      "grad_norm": 1.2283321619033813,
-      "learning_rate": 8e-05,
-      "loss": 0.8963,
+      "eval_loss": 1.063585877418518,
+      "eval_runtime": 281.615,
+      "eval_samples_per_second": 8.245,
+      "eval_steps_per_second": 8.245,
       "step": 4
     },
     {
       "epoch": 0.00045333998231974066,
-      "grad_norm": 1.0397660732269287,
+      "grad_norm": 1.351820945739746,
       "learning_rate": 0.0001,
-      "loss": 1.064,
+      "loss": 1.0148,
       "step": 5
     },
     {
       "epoch": 0.0005440079787836889,
-      "grad_norm": 1.3788177967071533,
+      "grad_norm": 1.6633617877960205,
       "learning_rate": 0.00012,
-      "loss": 1.311,
-      "step": 6
-    },
-    {
-      "epoch": 0.0005440079787836889,
-      "eval_loss": 0.9817929267883301,
-      "eval_runtime": 281.7359,
-      "eval_samples_per_second": 8.242,
-      "eval_steps_per_second": 8.242,
+      "loss": 1.1922,
       "step": 6
     },
     {
       "epoch": 0.000634675975247637,
-      "grad_norm": 1.0661911964416504,
+      "grad_norm": 1.216033697128296,
       "learning_rate": 0.00014,
-      "loss": 0.9209,
+      "loss": 0.7838,
       "step": 7
     },
     {
       "epoch": 0.0007253439717115851,
-      "grad_norm": 1.0511986017227173,
+      "grad_norm": 1.0886664390563965,
       "learning_rate": 0.00016,
-      "loss": 0.8409,
+      "loss": 0.6905,
       "step": 8
     },
     {
-      "epoch": 0.0008160119681755333,
-      "grad_norm": 1.0703239440917969,
-      "learning_rate": 0.00018,
-      "loss": 0.8303,
-      "step": 9
+      "epoch": 0.0007253439717115851,
+      "eval_loss": 0.6156781911849976,
+      "eval_runtime": 282.2824,
+      "eval_samples_per_second": 8.226,
+      "eval_steps_per_second": 8.226,
+      "step": 8
     },
     {
       "epoch": 0.0008160119681755333,
-      "eval_loss": 0.6860486268997192,
-      "eval_runtime": 285.0723,
-      "eval_samples_per_second": 8.145,
-      "eval_steps_per_second": 8.145,
+      "grad_norm": 1.2936314344406128,
+      "learning_rate": 0.00018,
+      "loss": 0.6603,
       "step": 9
     },
     {
       "epoch": 0.0009066799646394813,
-      "grad_norm": 1.1654818058013916,
+      "grad_norm": 1.429738163948059,
       "learning_rate": 0.0002,
-      "loss": 0.7193,
+      "loss": 0.518,
       "step": 10
     },
     {
       "epoch": 0.0009973479611034295,
-      "grad_norm": 1.2375497817993164,
-      "learning_rate": 0.00019781476007338058,
-      "loss": 0.6006,
+      "grad_norm": 1.501217007637024,
+      "learning_rate": 0.0001996917333733128,
+      "loss": 0.475,
       "step": 11
     },
     {
       "epoch": 0.0010880159575673777,
-      "grad_norm": 0.930781900882721,
-      "learning_rate": 0.0001913545457642601,
-      "loss": 0.4477,
+      "grad_norm": 1.0113286972045898,
+      "learning_rate": 0.00019876883405951377,
+      "loss": 0.3233,
       "step": 12
     },
     {
       "epoch": 0.0010880159575673777,
-      "eval_loss": 0.43009814620018005,
-      "eval_runtime": 283.6546,
-      "eval_samples_per_second": 8.186,
-      "eval_steps_per_second": 8.186,
+      "eval_loss": 0.29599475860595703,
+      "eval_runtime": 282.4723,
+      "eval_samples_per_second": 8.22,
+      "eval_steps_per_second": 8.22,
       "step": 12
     },
     {
       "epoch": 0.0011786839540313258,
-      "grad_norm": 1.0492234230041504,
-      "learning_rate": 0.00018090169943749476,
-      "loss": 0.4208,
+      "grad_norm": 1.045129656791687,
+      "learning_rate": 0.00019723699203976766,
+      "loss": 0.305,
       "step": 13
     },
     {
       "epoch": 0.001269351950495274,
-      "grad_norm": 0.8570551872253418,
-      "learning_rate": 0.00016691306063588583,
-      "loss": 0.3245,
+      "grad_norm": 0.7782332897186279,
+      "learning_rate": 0.00019510565162951537,
+      "loss": 0.2367,
       "step": 14
     },
     {
       "epoch": 0.001360019946959222,
-      "grad_norm": 1.2632145881652832,
-      "learning_rate": 0.00015000000000000001,
-      "loss": 0.2958,
+      "grad_norm": 1.1300699710845947,
+      "learning_rate": 0.0001923879532511287,
+      "loss": 0.1923,
       "step": 15
     },
     {
-      "epoch": 0.001360019946959222,
-      "eval_loss": 0.2870440185070038,
-      "eval_runtime": 282.0334,
-      "eval_samples_per_second": 8.233,
-      "eval_steps_per_second": 8.233,
-      "step": 15
+      "epoch": 0.0014506879434231703,
+      "grad_norm": 0.6594939231872559,
+      "learning_rate": 0.0001891006524188368,
+      "loss": 0.1823,
+      "step": 16
     },
     {
       "epoch": 0.0014506879434231703,
-      "grad_norm": 0.8736838698387146,
-      "learning_rate": 0.00013090169943749476,
-      "loss": 0.2332,
+      "eval_loss": 0.2610986828804016,
+      "eval_runtime": 282.2669,
+      "eval_samples_per_second": 8.226,
+      "eval_steps_per_second": 8.226,
       "step": 16
     },
     {
       "epoch": 0.0015413559398871183,
-      "grad_norm": 0.8168073892593384,
-      "learning_rate": 0.00011045284632676536,
-      "loss": 0.2588,
+      "grad_norm": 1.0713460445404053,
+      "learning_rate": 0.00018526401643540922,
+      "loss": 0.2457,
       "step": 17
     },
     {
       "epoch": 0.0016320239363510666,
-      "grad_norm": 0.9538082480430603,
-      "learning_rate": 8.954715367323468e-05,
-      "loss": 0.2894,
+      "grad_norm": 1.2344204187393188,
+      "learning_rate": 0.00018090169943749476,
+      "loss": 0.272,
       "step": 18
     },
     {
-      "epoch": 0.0016320239363510666,
-      "eval_loss": 0.2599656581878662,
-      "eval_runtime": 281.2105,
-      "eval_samples_per_second": 8.257,
-      "eval_steps_per_second": 8.257,
-      "step": 18
+      "epoch": 0.0017226919328150146,
+      "grad_norm": 0.7098881006240845,
+      "learning_rate": 0.0001760405965600031,
+      "loss": 0.225,
+      "step": 19
+    },
+    {
+      "epoch": 0.0018133599292789627,
+      "grad_norm": 0.8920848965644836,
+      "learning_rate": 0.00017071067811865476,
+      "loss": 0.1479,
+      "step": 20
+    },
+    {
+      "epoch": 0.0018133599292789627,
+      "eval_loss": 0.23455704748630524,
+      "eval_runtime": 282.1688,
+      "eval_samples_per_second": 8.229,
+      "eval_steps_per_second": 8.229,
+      "step": 20
+    },
+    {
+      "epoch": 0.001904027925742911,
+      "grad_norm": 0.5211899876594543,
+      "learning_rate": 0.00016494480483301836,
+      "loss": 0.1664,
+      "step": 21
+    },
+    {
+      "epoch": 0.001994695922206859,
+      "grad_norm": 0.6904815435409546,
+      "learning_rate": 0.00015877852522924732,
+      "loss": 0.2361,
+      "step": 22
+    },
+    {
+      "epoch": 0.002085363918670807,
+      "grad_norm": 0.6779294013977051,
+      "learning_rate": 0.0001522498564715949,
+      "loss": 0.1939,
+      "step": 23
+    },
+    {
+      "epoch": 0.0021760319151347554,
+      "grad_norm": 0.8565766215324402,
+      "learning_rate": 0.00014539904997395468,
+      "loss": 0.2695,
+      "step": 24
+    },
+    {
+      "epoch": 0.0021760319151347554,
+      "eval_loss": 0.22548052668571472,
+      "eval_runtime": 282.2744,
+      "eval_samples_per_second": 8.226,
+      "eval_steps_per_second": 8.226,
+      "step": 24
     }
   ],
   "logging_steps": 1,
-  "max_steps": 25,
+  "max_steps": 50,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
-  "save_steps": 3,
+  "save_steps": 4,
   "stateful_callbacks": {
     "TrainerControl": {
       "args": {
@@ -208,7 +250,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 6843506903482368.0,
+  "total_flos": 9088277063663616.0,
   "train_batch_size": 1,
   "trial_name": null,
   "trial_params": null

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bf5f0cf90aa3deddbb881f441c0237bff1fd72ac4c2cbf0e67c55438798ba345
+oid sha256:586972196f15dd32a3d87fb4e0aad2c5279b9a281041e5bf0b54d207daa83d8a
 size 6776