Auto upload 2026-08-06T19:33:59.037264 (part 2)

.gitattributes

@@ -62,3 +62,4 @@ wandb/run-20260806_191235-e1jitk7d/run-e1jitk7d.wandb filter=lfs diff=lfs merge=
 wandb/run-20260806_191734-z63xzj1k/run-z63xzj1k.wandb filter=lfs diff=lfs merge=lfs -text
 out/glu-linear_run/training_log.jsonl filter=lfs diff=lfs merge=lfs -text
 wandb/run-20260806_192925-br0wnn10/run-br0wnn10.wandb filter=lfs diff=lfs merge=lfs -text
+out/glu-situglu_run/training_log.jsonl filter=lfs diff=lfs merge=lfs -text

out/glu-silu_run/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e71210d8020fd558967728f9ccd64e39d8d92db7f811ae82645ddc20e5cd3bb0
+oid sha256:2f5719390447f244f1d64eb0ef7fe03f0a6969c13e43fd14b2f0863f3346c8bf
 size 31992144

out/glu-silu_run/training_log.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:13dc63198b634bcdc0fcbfb01956f1932e3f3561f33bff64ad405a18556984c3
-size 51647991
+oid sha256:ed459ea7ff0dc984b2ed06ca299f983f41b168e3a68d50d76cdf317699ac4ac9
+size 51650600

out/glu-situglu_run/checkpoint-500/config.json

@@ -18,7 +18,7 @@
   "mlp_type": "glu",
   "model_type": "tiny_llama",
   "num_attention_heads": 4,
-  "num_hidden_layers": 9,
+  "num_hidden_layers": 46,
   "num_key_value_heads": 4,
   "pad_token_id": 0,
   "pretraining_tp": 1,

out/glu-situglu_run/checkpoint-500/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ea2cacb7c490d4aee513a718c9d3355f576140e674a761d8cb41452c5f465d0c
-size 4011552
+oid sha256:cee31507219f8540f103315892624460c956636545177ee325ab7a536723af8f
+size 16191392

out/glu-situglu_run/checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:48b69466afc8f3131089b68950fdbdb0003bbde2bdd0cc8e1b0b3510a4bf86a3
-size 8074746
+oid sha256:e6f5316f255f2f86dfbf983a91300b711a1c4680cf1a00c70c63599edccf7eb7
+size 32639090

out/glu-situglu_run/checkpoint-500/trainer_state.json

@@ -13,191 +13,191 @@
       "epoch": 0.0013481631277384564,
       "grad_norm": 1.2578125,
       "learning_rate": 0.0003,
-      "loss": 7.793804931640625,
+      "loss": 7.777315521240235,
       "step": 20
     },
     {
       "epoch": 0.002696326255476913,
-      "grad_norm": 1.171875,
+      "grad_norm": 1.1640625,
       "learning_rate": 0.0003,
-      "loss": 7.022835540771484,
+      "loss": 7.018241119384766,
       "step": 40
     },
     {
       "epoch": 0.004044489383215369,
-      "grad_norm": 1.0,
+      "grad_norm": 0.9296875,
       "learning_rate": 0.0003,
-      "loss": 6.450706481933594,
+      "loss": 6.487100982666016,
       "step": 60
     },
     {
       "epoch": 0.005392652510953826,
-      "grad_norm": 1.9296875,
+      "grad_norm": 0.68359375,
       "learning_rate": 0.0003,
-      "loss": 6.003887176513672,
+      "loss": 6.16792106628418,
       "step": 80
     },
     {
       "epoch": 0.006740815638692282,
-      "grad_norm": 2.359375,
+      "grad_norm": 0.82421875,
       "learning_rate": 0.0003,
-      "loss": 5.44603271484375,
+      "loss": 5.984156799316406,
       "step": 100
     },
     {
       "epoch": 0.008088978766430738,
-      "grad_norm": 3.109375,
+      "grad_norm": 0.76171875,
       "learning_rate": 0.0003,
-      "loss": 4.916712951660156,
+      "loss": 5.7863212585449215,
       "step": 120
     },
     {
       "epoch": 0.009437141894169195,
-      "grad_norm": 3.609375,
+      "grad_norm": 0.8984375,
       "learning_rate": 0.0003,
-      "loss": 4.487728118896484,
+      "loss": 5.399650573730469,
       "step": 140
     },
     {
       "epoch": 0.010785305021907651,
-      "grad_norm": 2.6875,
+      "grad_norm": 1.15625,
       "learning_rate": 0.0003,
-      "loss": 4.115962219238281,
+      "loss": 4.998098373413086,
       "step": 160
     },
     {
       "epoch": 0.012133468149646108,
-      "grad_norm": 3.15625,
+      "grad_norm": 1.6015625,
       "learning_rate": 0.0003,
-      "loss": 3.767821502685547,
+      "loss": 4.63086051940918,
       "step": 180
     },
     {
       "epoch": 0.013481631277384564,
-      "grad_norm": 3.21875,
+      "grad_norm": 1.390625,
       "learning_rate": 0.0003,
-      "loss": 3.421479415893555,
+      "loss": 4.290957260131836,
       "step": 200
     },
     {
       "epoch": 0.013481631277384564,
-      "eval_loss": 3.2420618534088135,
-      "eval_runtime": 9.319,
-      "eval_samples_per_second": 1022.324,
-      "eval_steps_per_second": 15.989,
+      "eval_loss": 4.09553861618042,
+      "eval_runtime": 17.94,
+      "eval_samples_per_second": 531.049,
+      "eval_steps_per_second": 8.305,
       "step": 200
     },
     {
       "epoch": 0.01482979440512302,
-      "grad_norm": 2.8125,
+      "grad_norm": 1.7421875,
       "learning_rate": 0.0003,
-      "loss": 3.0867557525634766,
+      "loss": 3.9101463317871095,
       "step": 220
     },
     {
       "epoch": 0.016177957532861477,
-      "grad_norm": 3.375,
+      "grad_norm": 1.7109375,
       "learning_rate": 0.0003,
-      "loss": 2.7438758850097655,
+      "loss": 3.544460678100586,
       "step": 240
     },
     {
       "epoch": 0.01752612066059993,
-      "grad_norm": 3.015625,
+      "grad_norm": 1.6875,
       "learning_rate": 0.0003,
-      "loss": 2.4241146087646483,
+      "loss": 3.2432941436767577,
       "step": 260
     },
     {
       "epoch": 0.01887428378833839,
-      "grad_norm": 2.46875,
+      "grad_norm": 2.296875,
       "learning_rate": 0.0003,
-      "loss": 2.1026111602783204,
+      "loss": 2.9642810821533203,
       "step": 280
     },
     {
       "epoch": 0.020222446916076844,
-      "grad_norm": 1.3984375,
+      "grad_norm": 2.515625,
       "learning_rate": 0.0003,
-      "loss": 1.7901939392089843,
+      "loss": 2.689479637145996,
       "step": 300
     },
     {
       "epoch": 0.021570610043815303,
-      "grad_norm": 1.8046875,
+      "grad_norm": 1.65625,
       "learning_rate": 0.0003,
-      "loss": 1.5043752670288086,
+      "loss": 2.4316022872924803,
       "step": 320
     },
     {
       "epoch": 0.022918773171553757,
-      "grad_norm": 1.890625,
+      "grad_norm": 2.0,
       "learning_rate": 0.0003,
-      "loss": 1.2800609588623046,
+      "loss": 2.192088508605957,
       "step": 340
     },
     {
       "epoch": 0.024266936299292215,
-      "grad_norm": 1.6796875,
+      "grad_norm": 1.609375,
       "learning_rate": 0.0003,
-      "loss": 1.1210617065429687,
+      "loss": 1.9697820663452148,
       "step": 360
     },
     {
       "epoch": 0.02561509942703067,
-      "grad_norm": 1.3359375,
+      "grad_norm": 1.5234375,
       "learning_rate": 0.0003,
-      "loss": 0.9836687088012696,
+      "loss": 1.7674951553344727,
       "step": 380
     },
     {
       "epoch": 0.026963262554769128,
-      "grad_norm": 1.1171875,
+      "grad_norm": 1.6171875,
       "learning_rate": 0.0003,
-      "loss": 0.8616189002990723,
+      "loss": 1.5742257118225098,
       "step": 400
     },
     {
       "epoch": 0.026963262554769128,
-      "eval_loss": 0.8057775497436523,
-      "eval_runtime": 9.1733,
-      "eval_samples_per_second": 1038.555,
-      "eval_steps_per_second": 16.243,
+      "eval_loss": 1.4880807399749756,
+      "eval_runtime": 15.5417,
+      "eval_samples_per_second": 612.997,
+      "eval_steps_per_second": 9.587,
       "step": 400
     },
     {
       "epoch": 0.028311425682507583,
-      "grad_norm": 0.6640625,
+      "grad_norm": 1.265625,
       "learning_rate": 0.0003,
-      "loss": 0.744015121459961,
+      "loss": 1.401127243041992,
       "step": 420
     },
     {
       "epoch": 0.02965958881024604,
-      "grad_norm": 0.423828125,
+      "grad_norm": 1.2265625,
       "learning_rate": 0.0003,
-      "loss": 0.6658899784088135,
+      "loss": 1.2694280624389649,
       "step": 440
     },
     {
       "epoch": 0.031007751937984496,
-      "grad_norm": 0.37109375,
+      "grad_norm": 1.3671875,
       "learning_rate": 0.0003,
-      "loss": 0.5927415370941163,
+      "loss": 1.124942684173584,
       "step": 460
     },
     {
       "epoch": 0.032355915065722954,
-      "grad_norm": 0.41796875,
+      "grad_norm": 1.546875,
       "learning_rate": 0.0003,
-      "loss": 0.5481887817382812,
+      "loss": 1.0088685989379882,
       "step": 480
     },
     {
       "epoch": 0.03370407819346141,
-      "grad_norm": 0.376953125,
+      "grad_norm": 1.2109375,
       "learning_rate": 0.0003,
-      "loss": 0.5076817989349365,
+      "loss": 0.9041744232177734,
       "step": 500
     }
   ],
@@ -218,7 +218,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 145194221568000.0,
+  "total_flos": 742052069376000.0,
   "train_batch_size": 64,
   "trial_name": null,
   "trial_params": null

out/glu-situglu_run/checkpoint-500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ef27498240193108e38674f939b83e00bd410c944700e69c4b1c309cb945fd64
+oid sha256:6de53d54d65afa37f7c0133b97a7e72aa1e573b6271a5076e2f12640231686e5
 size 4856

out/glu-situglu_run/config.json

@@ -18,7 +18,7 @@
   "mlp_type": "glu",
   "model_type": "tiny_llama",
   "num_attention_heads": 4,
-  "num_hidden_layers": 9,
+  "num_hidden_layers": 46,
   "num_key_value_heads": 4,
   "pad_token_id": 0,
   "pretraining_tp": 1,

out/glu-situglu_run/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9cca2935db952e17a7b49c3e4b2d0557ad2f667b5568b7afe8f27712b3969a74
-size 4011552
+oid sha256:cee31507219f8540f103315892624460c956636545177ee325ab7a536723af8f
+size 16191392

out/glu-situglu_run/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ef27498240193108e38674f939b83e00bd410c944700e69c4b1c309cb945fd64
+oid sha256:6de53d54d65afa37f7c0133b97a7e72aa1e573b6271a5076e2f12640231686e5
 size 4856