Upload Qwen2ForCausalLM

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "final_run_qwen_0.5/run-0/checkpoint-8800",
+  "_name_or_path": "final_run_qwen_0.5/run-7/checkpoint-28993",
   "architectures": [
     "Qwen2ForCausalLM"
   ],
@@ -10,7 +10,7 @@
   "hidden_size": 896,
   "initializer_range": 0.02,
   "intermediate_size": 4864,
-  "max_position_embeddings": 131072,
+  "max_position_embeddings": 32768,
   "max_window_layers": 24,
   "model_type": "qwen2",
   "num_attention_heads": 14,
@@ -22,8 +22,9 @@
   "sliding_window": null,
   "tie_word_embeddings": true,
   "torch_dtype": "float16",
-  "transformers_version": "4.46.3",
+  "transformers_version": "4.47.1",
   "use_cache": true,
+  "use_mrope": false,
   "use_sliding_window": false,
   "vocab_size": 151936
 }

generation_config.json

@@ -2,5 +2,5 @@
   "bos_token_id": 151643,
   "eos_token_id": 151643,
   "max_new_tokens": 2048,
-  "transformers_version": "4.46.3"
+  "transformers_version": "4.47.1"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:720218e058f2c5a27af8881f0b653dab4b20134528ce16628db1e22d8f453bae
+oid sha256:d5280185916829f538fcaf78bd622c0d307a561384541bce21b72dd3d702d8b6
 size 988097536