Upload Qwen3ForCausalLM

config.json

@@ -4,9 +4,7 @@
   ],
   "attention_bias": false,
   "attention_dropout": 0.0,
-  "bos_token_id": 151643,
-  "dtype": "bfloat16",
-  "eos_token_id": 151645,
+  "dtype": "float32",
   "head_dim": 128,
   "hidden_act": "silu",
   "hidden_size": 8,
@@ -16,19 +14,19 @@
     "full_attention",
     "full_attention"
   ],
-  "max_position_embeddings": 40960,
-  "max_window_layers": 36,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
   "model_type": "qwen3",
   "num_attention_heads": 4,
   "num_hidden_layers": 2,
   "num_key_value_heads": 2,
   "rms_norm_eps": 1e-06,
   "rope_scaling": null,
-  "rope_theta": 1000000,
+  "rope_theta": 10000.0,
   "sliding_window": null,
   "tie_word_embeddings": false,
   "transformers_version": "4.57.0.dev0",
   "use_cache": true,
   "use_sliding_window": false,
-  "vocab_size": 151936
+  "vocab_size": 151669
 }

generation_config.json

@@ -1,6 +1,4 @@
 {
   "_from_model_config": true,
-  "bos_token_id": 151643,
-  "eos_token_id": 151645,
   "transformers_version": "4.57.0.dev0"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:22755aa79bbd9567aadd8966228132a0e6f8676a2aeaab4b208df7974cf8c116
-size 4917896
+oid sha256:a34fc1348710bb8151c80c5ae44b05c0fb7ab788f657e08160d57d4c5339eef7
+size 9816064