Upload task output 1

adapter_config.json

@@ -30,12 +30,12 @@
   "revision": null,
   "target_modules": [
     "gate_proj",
-    "v_proj",
-    "up_proj",
     "k_proj",
-    "q_proj",
+    "up_proj",
+    "v_proj",
+    "o_proj",
     "down_proj",
-    "o_proj"
+    "q_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5d5508f5dc8171e7c93d74369b8e1af48e33dffac0fbda8ae7e54838fea4911d
+oid sha256:f9f96b3def23987d813c156a82b958258856c9438ce56ed3b44c5e48ce68b916
 size 194563400

loss.txt

@@ -1 +1 @@
-9,no_eval
+1,no_eval

trainer_state.json

@@ -2,9 +2,9 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.00036,
+  "epoch": 4e-05,
   "eval_steps": 500,
-  "global_step": 9,
+  "global_step": 1,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -16,300 +16,36 @@
       "clip_ratio/low_min": 0.0,
       "clip_ratio/region_mean": 0.0,
       "completions/clipped_ratio": 0.0,
-      "completions/max_length": 1073.0,
-      "completions/max_terminated_length": 1073.0,
-      "completions/mean_length": 927.5625,
-      "completions/mean_terminated_length": 927.5625,
-      "completions/min_length": 289.0,
-      "completions/min_terminated_length": 289.0,
-      "entropy": 0.575973030179739,
+      "completions/max_length": 573.0,
+      "completions/max_terminated_length": 573.0,
+      "completions/mean_length": 125.71875,
+      "completions/mean_terminated_length": 125.71875,
+      "completions/min_length": 2.0,
+      "completions/min_terminated_length": 2.0,
+      "entropy": 1.372788105159998,
       "epoch": 4e-05,
       "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.9458226561546326,
+      "grad_norm": 0.7595316171646118,
       "kl": 0.0,
       "learning_rate": 0.0,
-      "loss": -0.0108,
-      "num_tokens": 129845.0,
-      "reward": -0.488802433013916,
-      "reward_std": 0.18876618146896362,
-      "rewards/rollout_reward_func/mean": -0.488802433013916,
-      "rewards/rollout_reward_func/std": 0.1920391470193863,
-      "sampling/importance_sampling_ratio/max": 1.2831777334213257,
-      "sampling/importance_sampling_ratio/mean": 1.0011134147644043,
-      "sampling/importance_sampling_ratio/min": 0.7999967932701111,
-      "sampling/sampling_logp_difference/max": 0.24933958053588867,
-      "sampling/sampling_logp_difference/mean": 0.01707676611840725,
+      "loss": -0.0345,
+      "num_tokens": 91750.0,
+      "reward": -0.6493574976921082,
+      "reward_std": 0.7037926316261292,
+      "rewards/rollout_reward_func/mean": -0.6493574976921082,
+      "rewards/rollout_reward_func/std": 0.8106608986854553,
+      "sampling/importance_sampling_ratio/max": 1.3807494640350342,
+      "sampling/importance_sampling_ratio/mean": 1.0163333415985107,
+      "sampling/importance_sampling_ratio/min": 0.7142292261123657,
+      "sampling/sampling_logp_difference/max": 0.23920273780822754,
+      "sampling/sampling_logp_difference/mean": 0.028075017035007477,
       "step": 1,
-      "step_time": 18.75738557600016
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": 0.0,
-      "completions/max_length": 1054.0,
-      "completions/max_terminated_length": 1054.0,
-      "completions/mean_length": 954.1875,
-      "completions/mean_terminated_length": 954.1875,
-      "completions/min_length": 691.0,
-      "completions/min_terminated_length": 691.0,
-      "entropy": 0.5630898363888264,
-      "epoch": 8e-05,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.8123396039009094,
-      "kl": 0.0,
-      "learning_rate": 2.8571428571428575e-07,
-      "loss": -0.0021,
-      "num_tokens": 261450.0,
-      "reward": -0.4951375424861908,
-      "reward_std": 0.18179327249526978,
-      "rewards/rollout_reward_func/mean": -0.4951375424861908,
-      "rewards/rollout_reward_func/std": 0.18824981153011322,
-      "sampling/importance_sampling_ratio/max": 1.2119427919387817,
-      "sampling/importance_sampling_ratio/mean": 1.002686858177185,
-      "sampling/importance_sampling_ratio/min": 0.8387882113456726,
-      "sampling/sampling_logp_difference/max": 0.19222474098205566,
-      "sampling/sampling_logp_difference/mean": 0.01698336750268936,
-      "step": 2,
-      "step_time": 18.354596549000235
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": 0.0,
-      "completions/max_length": 1039.0,
-      "completions/max_terminated_length": 1039.0,
-      "completions/mean_length": 948.046875,
-      "completions/mean_terminated_length": 948.046875,
-      "completions/min_length": 711.0,
-      "completions/min_terminated_length": 711.0,
-      "entropy": 0.567595299333334,
-      "epoch": 0.00012,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.7247381210327148,
-      "kl": 0.0005365689157770248,
-      "learning_rate": 5.714285714285715e-07,
-      "loss": 0.0017,
-      "num_tokens": 395514.0,
-      "reward": -0.5213372707366943,
-      "reward_std": 0.18358027935028076,
-      "rewards/rollout_reward_func/mean": -0.5213372707366943,
-      "rewards/rollout_reward_func/std": 0.1988830268383026,
-      "sampling/importance_sampling_ratio/max": 1.2515581846237183,
-      "sampling/importance_sampling_ratio/mean": 1.0002529621124268,
-      "sampling/importance_sampling_ratio/min": 0.8503696918487549,
-      "sampling/sampling_logp_difference/max": 0.22438931465148926,
-      "sampling/sampling_logp_difference/mean": 0.017911788076162338,
-      "step": 3,
-      "step_time": 21.38302714500037
-    },
-    {
-      "clip_ratio/high_max": 0.0062500000931322575,
-      "clip_ratio/high_mean": 0.0015625000232830644,
-      "clip_ratio/low_mean": 0.0015625000232830644,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0031250000465661287,
-      "completions/clipped_ratio": 0.0,
-      "completions/max_length": 1031.0,
-      "completions/max_terminated_length": 1031.0,
-      "completions/mean_length": 936.15625,
-      "completions/mean_terminated_length": 936.15625,
-      "completions/min_length": 653.0,
-      "completions/min_terminated_length": 653.0,
-      "entropy": 0.5737008973956108,
-      "epoch": 0.00016,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.7101178765296936,
-      "kl": 0.0005989186938677449,
-      "learning_rate": 8.571428571428572e-07,
-      "loss": -0.0103,
-      "num_tokens": 532318.0,
-      "reward": -0.4770505428314209,
-      "reward_std": 0.20475129783153534,
-      "rewards/rollout_reward_func/mean": -0.4770505130290985,
-      "rewards/rollout_reward_func/std": 0.21575571596622467,
-      "sampling/importance_sampling_ratio/max": 1.1857075691223145,
-      "sampling/importance_sampling_ratio/mean": 1.0007572174072266,
-      "sampling/importance_sampling_ratio/min": 0.7442525625228882,
-      "sampling/sampling_logp_difference/max": 0.29537487030029297,
-      "sampling/sampling_logp_difference/mean": 0.018634025007486343,
-      "step": 4,
-      "step_time": 21.295585246000428
-    },
-    {
-      "clip_ratio/high_max": 0.0052083334885537624,
-      "clip_ratio/high_mean": 0.0013020833721384406,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0013020833721384406,
-      "completions/clipped_ratio": 0.0,
-      "completions/max_length": 1354.0,
-      "completions/max_terminated_length": 1354.0,
-      "completions/mean_length": 1222.15625,
-      "completions/mean_terminated_length": 1222.15625,
-      "completions/min_length": 868.0,
-      "completions/min_terminated_length": 868.0,
-      "entropy": 0.5727464407682419,
-      "epoch": 0.0002,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.7757085561752319,
-      "kl": 0.0007553649911642424,
-      "learning_rate": 1.142857142857143e-06,
-      "loss": -0.0049,
-      "num_tokens": 685325.0,
-      "reward": -0.4957820177078247,
-      "reward_std": 0.15593528747558594,
-      "rewards/rollout_reward_func/mean": -0.4957820177078247,
-      "rewards/rollout_reward_func/std": 0.15906599164009094,
-      "sampling/importance_sampling_ratio/max": 1.3013360500335693,
-      "sampling/importance_sampling_ratio/mean": 0.9994803667068481,
-      "sampling/importance_sampling_ratio/min": 0.7946972846984863,
-      "sampling/sampling_logp_difference/max": 0.26339149475097656,
-      "sampling/sampling_logp_difference/mean": 0.017627151682972908,
-      "step": 5,
-      "step_time": 28.664589129999968
-    },
-    {
-      "clip_ratio/high_max": 0.0052083334885537624,
-      "clip_ratio/high_mean": 0.0013020833721384406,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0013020833721384406,
-      "completions/clipped_ratio": 0.0,
-      "completions/max_length": 1351.0,
-      "completions/max_terminated_length": 1351.0,
-      "completions/mean_length": 1219.46875,
-      "completions/mean_terminated_length": 1219.46875,
-      "completions/min_length": 1012.0,
-      "completions/min_terminated_length": 1012.0,
-      "entropy": 0.6081138551235199,
-      "epoch": 0.00024,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.6916493773460388,
-      "kl": 0.0005982896527712,
-      "learning_rate": 1.4285714285714286e-06,
-      "loss": -0.0088,
-      "num_tokens": 840964.0,
-      "reward": -0.5021334886550903,
-      "reward_std": 0.17595317959785461,
-      "rewards/rollout_reward_func/mean": -0.5021334886550903,
-      "rewards/rollout_reward_func/std": 0.17878760397434235,
-      "sampling/importance_sampling_ratio/max": 1.214083194732666,
-      "sampling/importance_sampling_ratio/mean": 1.0006656646728516,
-      "sampling/importance_sampling_ratio/min": 0.8280104994773865,
-      "sampling/sampling_logp_difference/max": 0.19398927688598633,
-      "sampling/sampling_logp_difference/mean": 0.018302714452147484,
-      "step": 6,
-      "step_time": 30.956883577000326
-    },
-    {
-      "clip_ratio/high_max": 0.0,
-      "clip_ratio/high_mean": 0.0,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0,
-      "completions/clipped_ratio": 0.015625,
-      "completions/max_length": 1358.0,
-      "completions/max_terminated_length": 1358.0,
-      "completions/mean_length": 1176.265625,
-      "completions/mean_terminated_length": 1175.71435546875,
-      "completions/min_length": 289.0,
-      "completions/min_terminated_length": 289.0,
-      "entropy": 0.5972259026020765,
-      "epoch": 0.00028,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.7177211046218872,
-      "kl": 0.0005861011686647544,
-      "learning_rate": 1.7142857142857145e-06,
-      "loss": -0.0394,
-      "num_tokens": 996609.0,
-      "reward": -0.492123544216156,
-      "reward_std": 0.16847285628318787,
-      "rewards/rollout_reward_func/mean": -0.492123544216156,
-      "rewards/rollout_reward_func/std": 0.18208079040050507,
-      "sampling/importance_sampling_ratio/max": 1.2812676429748535,
-      "sampling/importance_sampling_ratio/mean": 0.9978591203689575,
-      "sampling/importance_sampling_ratio/min": 0.7220999002456665,
-      "sampling/sampling_logp_difference/max": 0.3255918025970459,
-      "sampling/sampling_logp_difference/mean": 0.017380019649863243,
-      "step": 7,
-      "step_time": 32.67134758100042
-    },
-    {
-      "clip_ratio/high_max": 0.0052083334885537624,
-      "clip_ratio/high_mean": 0.0013020833721384406,
-      "clip_ratio/low_mean": 0.0,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0013020833721384406,
-      "completions/clipped_ratio": 0.0,
-      "completions/max_length": 1358.0,
-      "completions/max_terminated_length": 1358.0,
-      "completions/mean_length": 1221.671875,
-      "completions/mean_terminated_length": 1221.671875,
-      "completions/min_length": 790.0,
-      "completions/min_terminated_length": 790.0,
-      "entropy": 0.6069452427327633,
-      "epoch": 0.00032,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.7911893129348755,
-      "kl": 0.0007064663786877645,
-      "learning_rate": 2.0000000000000003e-06,
-      "loss": 0.0074,
-      "num_tokens": 1151043.0,
-      "reward": -0.5468676686286926,
-      "reward_std": 0.17733421921730042,
-      "rewards/rollout_reward_func/mean": -0.5468676686286926,
-      "rewards/rollout_reward_func/std": 0.1824226826429367,
-      "sampling/importance_sampling_ratio/max": 1.1498843431472778,
-      "sampling/importance_sampling_ratio/mean": 0.9998562335968018,
-      "sampling/importance_sampling_ratio/min": 0.7584721446037292,
-      "sampling/sampling_logp_difference/max": 0.27644920349121094,
-      "sampling/sampling_logp_difference/mean": 0.01727226749062538,
-      "step": 8,
-      "step_time": 33.65800914100009
-    },
-    {
-      "clip_ratio/high_max": 0.004464285913854837,
-      "clip_ratio/high_mean": 0.0011160714784637094,
-      "clip_ratio/low_mean": 0.0011160714784637094,
-      "clip_ratio/low_min": 0.0,
-      "clip_ratio/region_mean": 0.0022321429569274187,
-      "completions/clipped_ratio": 0.0,
-      "completions/max_length": 1576.0,
-      "completions/max_terminated_length": 1576.0,
-      "completions/mean_length": 1432.5625,
-      "completions/mean_terminated_length": 1432.5625,
-      "completions/min_length": 1167.0,
-      "completions/min_terminated_length": 1167.0,
-      "entropy": 0.6040426194667816,
-      "epoch": 0.00036,
-      "frac_reward_zero_std": 0.0,
-      "grad_norm": 0.9153677821159363,
-      "kl": 0.0006420876543415943,
-      "learning_rate": 2.285714285714286e-06,
-      "loss": 0.0259,
-      "num_tokens": 1321059.0,
-      "reward": -0.5375348925590515,
-      "reward_std": 0.19819244742393494,
-      "rewards/rollout_reward_func/mean": -0.5375348925590515,
-      "rewards/rollout_reward_func/std": 0.20910826325416565,
-      "sampling/importance_sampling_ratio/max": 1.264775037765503,
-      "sampling/importance_sampling_ratio/mean": 1.0003395080566406,
-      "sampling/importance_sampling_ratio/min": 0.7903153896331787,
-      "sampling/sampling_logp_difference/max": 0.23532319068908691,
-      "sampling/sampling_logp_difference/mean": 0.01806110143661499,
-      "step": 9,
-      "step_time": 38.66699894299927
+      "step_time": 10.923513879000893
     }
   ],
   "logging_steps": 1.0,
-  "max_steps": 600,
-  "num_input_tokens_seen": 1321059,
+  "max_steps": 1800,
+  "num_input_tokens_seen": 91750,
   "num_train_epochs": 1,
   "save_steps": 500,
   "stateful_callbacks": {

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8421f1a93a2c8cf9dd383121cf9ef8942a208c5e26428c5eb482d90b23265824
+oid sha256:64e7205f90417143f6f991cb0e59138a7957c08b7eabdb09fcfc8d2e57680109
 size 8145