Upload MistralForSequenceClassification

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "reranker-malaysian-mistral-191M/checkpoint-27200",
+  "_name_or_path": "reranker-malaysian-mistral-349M/checkpoint-22800",
   "architectures": [
     "MistralForSequenceClassification"
   ],
@@ -9,15 +9,15 @@
   },
   "bos_token_id": 1,
   "eos_token_id": 2,
-  "head_dim": 48,
+  "head_dim": 64,
   "hidden_act": "silu",
-  "hidden_size": 768,
+  "hidden_size": 1024,
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
+  "intermediate_size": 4096,
   "max_position_embeddings": 4096,
   "model_type": "mistral",
   "num_attention_heads": 16,
-  "num_hidden_layers": 16,
+  "num_hidden_layers": 18,
   "num_key_value_heads": 8,
   "pad_token_id": 0,
   "problem_type": "single_label_classification",
@@ -26,7 +26,7 @@
   "sliding_window": 4096,
   "tie_word_embeddings": false,
   "torch_dtype": "float32",
-  "transformers_version": "4.44.2",
+  "transformers_version": "4.43.3",
   "use_cache": true,
   "vocab_size": 32000
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1b1fa49d61d8c411173d7eb130b0fd4d9baa56e9945d69014d11ff07906f52c8
-size 664658904
+oid sha256:1e21e652881462efabad0555df109f3abfc990a46afdeefef60f3fe542cdb486
+size 1263712472