latest = step 1500 (nDCG@10=0.4498)

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:96bed2cdda97dd7cf15f58cb353885069b4954e42704150ac67f4320d582dcf9
+oid sha256:1c256b3723c0993bf2ee89be89e5a21231692159d44e6bc93d7a719fbf16924b
 size 280019040

trainer_state.json

@@ -1,10 +1,10 @@
 {
-  "best_global_step": 1000,
-  "best_metric": 0.4370587039801932,
-  "best_model_checkpoint": "checkpoints/med-ruri-v3-70m-v2-from-med/checkpoint-1000",
-  "epoch": 0.1623640201331385,
+  "best_global_step": 1500,
+  "best_metric": 0.4497772162074601,
+  "best_model_checkpoint": "checkpoints/med-ruri-v3-70m-v2-from-med/checkpoint-1500",
+  "epoch": 0.24354603019970775,
   "eval_steps": 500,
-  "global_step": 1000,
+  "global_step": 1500,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -186,6 +186,95 @@
       "eval_samples_per_second": 0.0,
       "eval_steps_per_second": 0.0,
       "step": 1000
+    },
+    {
+      "epoch": 0.17048222113979541,
+      "grad_norm": 23.29427146911621,
+      "learning_rate": 8.733549820543497e-06,
+      "loss": 0.6038346862792969,
+      "step": 1050
+    },
+    {
+      "epoch": 0.17860042214645236,
+      "grad_norm": 28.32552719116211,
+      "learning_rate": 8.648094342847378e-06,
+      "loss": 0.6003586959838867,
+      "step": 1100
+    },
+    {
+      "epoch": 0.18671862315310928,
+      "grad_norm": 30.75159454345703,
+      "learning_rate": 8.562638865151257e-06,
+      "loss": 0.6092750930786133,
+      "step": 1150
+    },
+    {
+      "epoch": 0.1948368241597662,
+      "grad_norm": 23.42568016052246,
+      "learning_rate": 8.477183387455136e-06,
+      "loss": 0.5733106994628906,
+      "step": 1200
+    },
+    {
+      "epoch": 0.2029550251664231,
+      "grad_norm": 26.38534927368164,
+      "learning_rate": 8.391727909759017e-06,
+      "loss": 0.6007821655273438,
+      "step": 1250
+    },
+    {
+      "epoch": 0.21107322617308005,
+      "grad_norm": 29.11765480041504,
+      "learning_rate": 8.306272432062896e-06,
+      "loss": 0.575584602355957,
+      "step": 1300
+    },
+    {
+      "epoch": 0.21919142717973697,
+      "grad_norm": 26.75505256652832,
+      "learning_rate": 8.220816954366777e-06,
+      "loss": 0.5628342819213867,
+      "step": 1350
+    },
+    {
+      "epoch": 0.22730962818639389,
+      "grad_norm": 26.10503387451172,
+      "learning_rate": 8.135361476670656e-06,
+      "loss": 0.5451336669921875,
+      "step": 1400
+    },
+    {
+      "epoch": 0.23542782919305083,
+      "grad_norm": 28.066715240478516,
+      "learning_rate": 8.049905998974535e-06,
+      "loss": 0.5360281372070312,
+      "step": 1450
+    },
+    {
+      "epoch": 0.24354603019970775,
+      "grad_norm": 26.54391860961914,
+      "learning_rate": 7.964450521278416e-06,
+      "loss": 0.5346591949462891,
+      "step": 1500
+    },
+    {
+      "epoch": 0.24354603019970775,
+      "eval_med-ir_cosine_accuracy@1": 0.2595,
+      "eval_med-ir_cosine_accuracy@10": 0.6725,
+      "eval_med-ir_cosine_accuracy@5": 0.55,
+      "eval_med-ir_cosine_map@10": 0.38029999999999997,
+      "eval_med-ir_cosine_mrr@10": 0.3803000000000005,
+      "eval_med-ir_cosine_ndcg@10": 0.4497772162074601,
+      "eval_med-ir_cosine_precision@1": 0.2595,
+      "eval_med-ir_cosine_precision@10": 0.06724999999999999,
+      "eval_med-ir_cosine_precision@5": 0.11,
+      "eval_med-ir_cosine_recall@1": 0.2595,
+      "eval_med-ir_cosine_recall@10": 0.6725,
+      "eval_med-ir_cosine_recall@5": 0.55,
+      "eval_runtime": 54.4304,
+      "eval_samples_per_second": 0.0,
+      "eval_steps_per_second": 0.0,
+      "step": 1500
     }
   ],
   "logging_steps": 50,