2890

.gitattributes

@@ -29,3 +29,5 @@ global_step2550/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=l
 global_step2550/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text
 global_step2720/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
 global_step2720/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step2890/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt filter=lfs diff=lfs merge=lfs -text
+global_step2890/mp_rank_00_model_states.pt filter=lfs diff=lfs merge=lfs -text

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1377a547a61df3e253fae7fdc28de08fb751c6561735db32918f4c8d525b0ab2
+oid sha256:a8d0d9804232b2468e167718a32e5d1469a348f40813e1235373edaf94d06ab8
 size 335606536

{global_step2720 → global_step2890}/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f4ace3614c23e2628738aeb127757a3c39cbbe5a76d5f7197f83af4a5c8f0bd8
+oid sha256:5ddf07eea9da9f99c6e2b08765d63990e5cb2bc9772cb43c58451bc6f037ac9a
 size 2013320965

{global_step2720 → global_step2890}/mp_rank_00_model_states.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f9fe347faa8bd24b2b578f94d3f2b4d1e8588c03109978f844f6d3d69270d193
+oid sha256:da88d15d65bf593fcf36ccb4865d48e8f9b42e8a4453a7c49e14000e93d5531a
 size 335901529

latest

@@ -1 +1 @@
-global_step2720
+global_step2890

trainer_state.json

@@ -1,10 +1,10 @@
 {
-  "best_global_step": 2550,
-  "best_metric": 0.3034323453903198,
-  "best_model_checkpoint": "/home/longtail/data/outputs/exaone_7b_lora_rtlcoder_ratio1.0/checkpoint-2550",
-  "epoch": 2.0497456189937817,
+  "best_global_step": 2890,
+  "best_metric": 0.3028174340724945,
+  "best_model_checkpoint": "/home/longtail/data/outputs/exaone_7b_lora_rtlcoder_ratio1.0/checkpoint-2890",
+  "epoch": 2.177878273977765,
   "eval_steps": 170,
-  "global_step": 2720,
+  "global_step": 2890,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -2904,6 +2904,187 @@
       "eval_samples_per_second": 5.963,
       "eval_steps_per_second": 1.492,
       "step": 2720
+    },
+    {
+      "entropy": 0.25732421875,
+      "epoch": 2.0572828339928395,
+      "grad_norm": 0.4523007571697235,
+      "learning_rate": 1.3298346751826624e-05,
+      "loss": 0.2659280300140381,
+      "mean_token_accuracy": 0.9151400104165077,
+      "num_tokens": 34574314.0,
+      "step": 2730
+    },
+    {
+      "entropy": 0.2736328125,
+      "epoch": 2.0648200489918973,
+      "grad_norm": 0.43264201283454895,
+      "learning_rate": 1.3252219281872663e-05,
+      "loss": 0.27533228397369386,
+      "mean_token_accuracy": 0.9147139310836792,
+      "num_tokens": 34694661.0,
+      "step": 2740
+    },
+    {
+      "entropy": 0.25009765625,
+      "epoch": 2.0723572639909555,
+      "grad_norm": 0.38429099321365356,
+      "learning_rate": 1.3206014297659538e-05,
+      "loss": 0.25780477523803713,
+      "mean_token_accuracy": 0.9180345565080643,
+      "num_tokens": 34821985.0,
+      "step": 2750
+    },
+    {
+      "entropy": 0.2557861328125,
+      "epoch": 2.0798944789900133,
+      "grad_norm": 0.40265947580337524,
+      "learning_rate": 1.315973290044913e-05,
+      "loss": 0.26505556106567385,
+      "mean_token_accuracy": 0.9161929801106453,
+      "num_tokens": 34946746.0,
+      "step": 2760
+    },
+    {
+      "entropy": 0.274853515625,
+      "epoch": 2.087431693989071,
+      "grad_norm": 0.3728543519973755,
+      "learning_rate": 1.3113376193324564e-05,
+      "loss": 0.2775618314743042,
+      "mean_token_accuracy": 0.9146954879164696,
+      "num_tokens": 35084683.0,
+      "step": 2770
+    },
+    {
+      "entropy": 0.26063232421875,
+      "epoch": 2.094968908988129,
+      "grad_norm": 0.4594137668609619,
+      "learning_rate": 1.3066945281163923e-05,
+      "loss": 0.2669402837753296,
+      "mean_token_accuracy": 0.9168649092316628,
+      "num_tokens": 35207447.0,
+      "step": 2780
+    },
+    {
+      "entropy": 0.267236328125,
+      "epoch": 2.1025061239871867,
+      "grad_norm": 0.4985879063606262,
+      "learning_rate": 1.302044127061392e-05,
+      "loss": 0.2751408815383911,
+      "mean_token_accuracy": 0.912787701189518,
+      "num_tokens": 35325558.0,
+      "step": 2790
+    },
+    {
+      "entropy": 0.26943359375,
+      "epoch": 2.1100433389862445,
+      "grad_norm": 0.45133885741233826,
+      "learning_rate": 1.2973865270063501e-05,
+      "loss": 0.27563629150390623,
+      "mean_token_accuracy": 0.9126975074410438,
+      "num_tokens": 35451140.0,
+      "step": 2800
+    },
+    {
+      "entropy": 0.2789306640625,
+      "epoch": 2.1175805539853023,
+      "grad_norm": 0.42350032925605774,
+      "learning_rate": 1.2927218389617452e-05,
+      "loss": 0.28734893798828126,
+      "mean_token_accuracy": 0.911960031092167,
+      "num_tokens": 35567497.0,
+      "step": 2810
+    },
+    {
+      "entropy": 0.255517578125,
+      "epoch": 2.12511776898436,
+      "grad_norm": 0.4202309548854828,
+      "learning_rate": 1.2880501741069931e-05,
+      "loss": 0.2654577255249023,
+      "mean_token_accuracy": 0.9167904317378998,
+      "num_tokens": 35695050.0,
+      "step": 2820
+    },
+    {
+      "entropy": 0.23223876953125,
+      "epoch": 2.132654983983418,
+      "grad_norm": 0.44368067383766174,
+      "learning_rate": 1.2833716437877952e-05,
+      "loss": 0.23806145191192626,
+      "mean_token_accuracy": 0.9233140185475349,
+      "num_tokens": 35822236.0,
+      "step": 2830
+    },
+    {
+      "entropy": 0.2786865234375,
+      "epoch": 2.140192198982476,
+      "grad_norm": 0.40076377987861633,
+      "learning_rate": 1.278686359513488e-05,
+      "loss": 0.2845803737640381,
+      "mean_token_accuracy": 0.9122401431202889,
+      "num_tokens": 35958799.0,
+      "step": 2840
+    },
+    {
+      "entropy": 0.2666748046875,
+      "epoch": 2.147729413981534,
+      "grad_norm": 0.4162449538707733,
+      "learning_rate": 1.2739944329543818e-05,
+      "loss": 0.274515962600708,
+      "mean_token_accuracy": 0.9145516887307167,
+      "num_tokens": 36083078.0,
+      "step": 2850
+    },
+    {
+      "entropy": 0.252099609375,
+      "epoch": 2.1552666289805917,
+      "grad_norm": 0.41252174973487854,
+      "learning_rate": 1.2692959759391027e-05,
+      "loss": 0.2631190299987793,
+      "mean_token_accuracy": 0.9182503208518028,
+      "num_tokens": 36203301.0,
+      "step": 2860
+    },
+    {
+      "entropy": 0.2442626953125,
+      "epoch": 2.1628038439796495,
+      "grad_norm": 0.47404617071151733,
+      "learning_rate": 1.2645911004519246e-05,
+      "loss": 0.250153923034668,
+      "mean_token_accuracy": 0.9218041867017746,
+      "num_tokens": 36337562.0,
+      "step": 2870
+    },
+    {
+      "entropy": 0.2638671875,
+      "epoch": 2.1703410589787073,
+      "grad_norm": 0.44618526101112366,
+      "learning_rate": 1.2598799186301003e-05,
+      "loss": 0.26806776523590087,
+      "mean_token_accuracy": 0.9138118460774421,
+      "num_tokens": 36451609.0,
+      "step": 2880
+    },
+    {
+      "entropy": 0.2649169921875,
+      "epoch": 2.177878273977765,
+      "grad_norm": 0.4570532441139221,
+      "learning_rate": 1.2551625427611907e-05,
+      "loss": 0.27152099609375,
+      "mean_token_accuracy": 0.9154557690024376,
+      "num_tokens": 36578083.0,
+      "step": 2890
+    },
+    {
+      "epoch": 2.177878273977765,
+      "eval_entropy": 0.2681899472891566,
+      "eval_loss": 0.3028174340724945,
+      "eval_mean_token_accuracy": 0.9079541289303676,
+      "eval_num_tokens": 36578083.0,
+      "eval_runtime": 445.3242,
+      "eval_samples_per_second": 5.957,
+      "eval_steps_per_second": 1.491,
+      "step": 2890
     }
   ],
   "logging_steps": 10,
@@ -2923,7 +3104,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 2.6097961573663375e+18,
+  "total_flos": 2.7708789068765594e+18,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null