Joint training checkpoint step 1523

README.md

@@ -10,7 +10,7 @@ tags: [tabular-data, masked-lm, token-classification, experimental]
 Experimental budget-limited mmBERT-base joint fine-tuning checkpoint.
 
 One shared encoder, 18 category BIO outputs, insertion outputs, and the pretrained MLM vocabulary head. Correction is trained in place: the selected corrupted span is replaced by a variable length mask buffer inside its original cell, with no appended JSON request. The MLM head emits replacement tokens followed by `END_EDIT` and visible padding tokens.
-Completed optimizer updates: **1400**. This checkpoint is a research baseline, not a validated automatic cleaning system.
+Completed optimizer updates: **1523**. This checkpoint is a research baseline, not a validated automatic cleaning system.
 
 ## Use and resume
 
@@ -2358,16 +2358,17 @@ The learning rate warms up for 100 updates, then follows cosine decay to 10% of
     "best_step": 800,
     "bad_checks": 3
   },
-  "last_conditional_pair_loss": 4.7715508117107674e-05,
+  "last_conditional_pair_loss": 7.90479316492565e-05,
   "peak_gpu_bytes": 7446971392,
-  "step": 1400,
-  "learning_rate": 2.7857757133684754e-06,
-  "last_detection_loss": 0.0021373224444687366,
-  "last_correction_loss": 0.1087012067437172,
-  "training_elapsed_seconds": 1925.0292960060005,
+  "step": 1523,
+  "learning_rate": 2.119826502129188e-06,
+  "last_detection_loss": 0.001190010691061616,
+  "last_correction_loss": 2.6506691938266158e-05,
+  "training_elapsed_seconds": 2101.1284097140006,
   "exclusions": {},
-  "mean_detection_loss": 0.12874416158301755,
-  "mean_correction_loss": 0.1242886986972735,
-  "loss_interval_updates": 10
+  "mean_detection_loss": 0.022690351656638085,
+  "mean_correction_loss": 0.12669519676537674,
+  "loss_interval_updates": 10,
+  "stop_reason": "time_budget"
 }
 ```

checkpoint.json

@@ -1,19 +1,19 @@
 {
   "version": 1,
-  "step": 1400,
+  "step": 1523,
   "dataset_revision": "7da5aa80bc2a68a375930f3bd3ed2f7fdda0152f",
   "base_revision": "c5955035435e2bf121cde7f3c8863ef52ff35d82",
   "files": {
-    "model.safetensors": "bdf79b7c0640babf10a7abaf8ed211d6c1ac1d0c66e9637fdc95353041d33f63",
+    "model.safetensors": "ccc60991dec5597d23fe1b2fa2f6504979cbd99f307f87cd85380b4c30a2a404",
     "config.json": "9e105ce4ed25ed2d6be2443eeb69932a530668a248ce3211d43a89a136d42531",
     "tokenizer_config.json": "51d83f5e4c1174f8dad62610076998730a57a9ac7025680653c720bca2a9b0a1",
     "tokenizer.json": "47e408472c05385fea5fb9f8b02bbd838d55da8f15a22c863a99fd31d89243f5",
-    "training_state.pt": "9574b53582e86f191d07ddee32914492867f8506ac5aad6d21e3e22f98d01515",
+    "training_state.pt": "1bbd4cbb4dc7ebb6d84b60419d2dd618e4d88b081c8986583cb516ed624a9909",
     "run_config.json": "eec4d8db620354de7883275f2c7282ee3bdd72328c3ff829001e248aeea49962",
-    "metrics.json": "942c2c676ec1ff5b0397a4be426c12e90c9ac87722160dc112b82ef4d0597b5c",
+    "metrics.json": "52be83173a7216364601332719e91e1ba11f9f9218f552c13a5d2772e6f875f1",
     "categories.json": "d95af1bcb70bc78d53297d4e57909776773fff76329098ece751e0734f03b29b",
     "train_tabfix.py": "8712d7afc651f2d3c55210f3a062ad6a19a94b6185b60d8b5593cacbdf31f7be",
-    "README.md": "9bcea147229642da47fdd78057b6c55035c8a71ff09647453452c3bf53d69267",
+    "README.md": "8e44c0726ce9ce09b9a26b7494779fc035f9baafe5d64b15a464989fcfa2b9ac",
     "LICENSE": "b6c1c7bb5aae03bb6b3396fc7c65cbe7306c422ed424b0b7b687d9a50cc15412",
     "BASE_MODEL_CARD.md": "35724415037ee3159aa733c16050620cc19a76219112a6ceeb3949572e0b37e9"
   },

metrics.json

@@ -2321,15 +2321,16 @@
     "best_step": 800,
     "bad_checks": 3
   },
-  "last_conditional_pair_loss": 4.7715508117107674e-05,
+  "last_conditional_pair_loss": 7.90479316492565e-05,
   "peak_gpu_bytes": 7446971392,
-  "step": 1400,
-  "learning_rate": 2.7857757133684754e-06,
-  "last_detection_loss": 0.0021373224444687366,
-  "last_correction_loss": 0.1087012067437172,
-  "training_elapsed_seconds": 1925.0292960060005,
+  "step": 1523,
+  "learning_rate": 2.119826502129188e-06,
+  "last_detection_loss": 0.001190010691061616,
+  "last_correction_loss": 2.6506691938266158e-05,
+  "training_elapsed_seconds": 2101.1284097140006,
   "exclusions": {},
-  "mean_detection_loss": 0.12874416158301755,
-  "mean_correction_loss": 0.1242886986972735,
-  "loss_interval_updates": 10
+  "mean_detection_loss": 0.022690351656638085,
+  "mean_correction_loss": 0.12669519676537674,
+  "loss_interval_updates": 10,
+  "stop_reason": "time_budget"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bdf79b7c0640babf10a7abaf8ed211d6c1ac1d0c66e9637fdc95353041d33f63
+oid sha256:ccc60991dec5597d23fe1b2fa2f6504979cbd99f307f87cd85380b4c30a2a404
 size 1231390724

training_state.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9574b53582e86f191d07ddee32914492867f8506ac5aad6d21e3e22f98d01515
-size 2462881846
+oid sha256:1bbd4cbb4dc7ebb6d84b60419d2dd618e4d88b081c8986583cb516ed624a9909
+size 2462881910