rotary_emb

config.json

@@ -5,13 +5,14 @@
     "Zamba2ForCausalLM"
   ],
   "attention_dropout": 0.0,
-  "attention_head_dim": 1,
+  "attention_head_dim": 2,
   "attention_hidden_size": 32,
   "bos_token_id": 1,
   "chunk_size": 256,
   "conv_kernel": 3,
   "d_model": 16,
   "d_state": 32,
+  "dtype": "float32",
   "eos_token_id": 2,
   "expand": 2,
   "hidden_act": "gelu",
@@ -36,9 +37,9 @@
   "max_position_embeddings": 4096,
   "model_type": "zamba2",
   "n_mamba_heads": 8,
-  "num_attention_heads": 32,
+  "num_attention_heads": 16,
   "num_hidden_layers": 4,
-  "num_key_value_heads": 32,
+  "num_key_value_heads": 16,
   "num_logits_to_keep": 1,
   "num_mem_blocks": 1,
   "num_query_groups": 32,
@@ -49,13 +50,12 @@
   "time_step_limit": null,
   "time_step_max": 0.1,
   "time_step_min": 0.001,
-  "torch_dtype": "float32",
-  "transformers_version": "4.55.4",
+  "transformers_version": "4.57.6",
   "use_cache": true,
   "use_conv_bias": true,
   "use_long_context": false,
   "use_mem_eff_path": false,
-  "use_mem_rope": false,
+  "use_mem_rope": true,
   "use_shared_attention_adapter": false,
   "vocab_size": 50280
 }

generation_config.json

@@ -3,5 +3,5 @@
   "bos_token_id": 1,
   "eos_token_id": 2,
   "pad_token_id": 0,
-  "transformers_version": "4.55.4"
+  "transformers_version": "4.57.6"
 }

pytorch_model.bin → model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fa28ee93e29e40c5fc35e5727122220d5dc937b65f27f71995c8edbb056e551c
-size 3408858
+oid sha256:245acf626383bc1f49e9e1630edb9d7ad3d2c3be2affda6db9978dc14ec38c8b
+size 3398168