Trained with Unsloth

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "unsloth/DeepSeek-R1-Distill-Qwen-7B-unsloth-bnb-4bit",
+  "_name_or_path": "agentica-org/DeepScaleR-1.5B-Preview",
   "architectures": [
     "Qwen2ForCausalLM"
   ],
@@ -7,16 +7,16 @@
   "bos_token_id": 151646,
   "eos_token_id": 151643,
   "hidden_act": "silu",
-  "hidden_size": 3584,
+  "hidden_size": 1536,
   "initializer_range": 0.02,
-  "intermediate_size": 18944,
+  "intermediate_size": 8960,
   "max_position_embeddings": 131072,
-  "max_window_layers": 28,
+  "max_window_layers": 21,
   "model_type": "qwen2",
-  "num_attention_heads": 28,
+  "num_attention_heads": 12,
   "num_hidden_layers": 28,
-  "num_key_value_heads": 4,
-  "pad_token_id": 151654,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
   "rms_norm_eps": 1e-06,
   "rope_scaling": null,
   "rope_theta": 10000,
@@ -24,10 +24,9 @@
   "tie_word_embeddings": false,
   "torch_dtype": "float16",
   "transformers_version": "4.48.2",
-  "unsloth_fixed": true,
   "unsloth_version": "2025.2.4",
   "use_cache": true,
   "use_mrope": false,
   "use_sliding_window": false,
-  "vocab_size": 152064
+  "vocab_size": 151936
 }

generation_config.json

@@ -4,7 +4,6 @@
   "do_sample": true,
   "eos_token_id": 151643,
   "max_length": 131072,
-  "pad_token_id": 151654,
   "temperature": 0.6,
   "top_p": 0.95,
   "transformers_version": "4.48.2"