Push model using huggingface_hub.

glirel_config.json

@@ -1,32 +1,39 @@
 {
   "lr_encoder": "1e-5",
   "lr_others": "1e-4",
-  "num_steps": 10000,
+  "num_steps": 300000,
   "warmup_ratio": 0.1,
-  "train_batch_size": 8,
-  "eval_every": 500,
-  "gradient_accumulation": null,
+  "train_batch_size": 1,
+  "eval_every": 3000,
+  "gradient_accumulation": 8,
   "eval_batch_size": 32,
   "max_width": 12,
   "model_name": "microsoft/deberta-v3-large",
   "fine_tune": true,
   "subtoken_pooling": "first",
   "hidden_size": 768,
+  "scorer": "dot",
   "span_mode": "marker",
+  "refine_prompt": false,
+  "refine_relation": false,
+  "ffn_mul": 4,
   "dropout": 0.4,
   "scheduler": "cosine_with_warmup",
   "loss_func": "binary_cross_entropy_loss",
-  "dataset_name": "few_rel",
+  "dataset_name": "zero_rel",
   "root_dir": "ablation_backbone",
-  "train_data": "data/few_rel_all.jsonl",
+  "train_data": [
+    "data/zero_rel_all.jsonl"
+  ],
   "prev_path": "none",
   "size_sup": -1,
   "num_train_rel_types": 25,
   "num_unseen_rel_types": 15,
   "top_k": 1,
-  "random_drop": true,
+  "random_drop": false,
   "max_len": 384,
-  "eval_threshold": 0.1,
+  "eval_threshold": 0.001,
   "name": "large",
-  "log_dir": null
+  "log_dir": "logs/zero_rel/zero_rel-2024-06-19__16-34-15",
+  "eval_data": "data/wiki_zsl_all.jsonl"
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ba1641e51dcc0123bbd26b4ea8b9946fdf5ad2d30c6c048ccc92dbfb069459e8
+oid sha256:296483cf26b60796cef031e0c53145826e18536233469fba0a03631f1da471b9
 size 1866500606