Upload LlamaForCausalLMEagle3

config.json

@@ -4,25 +4,35 @@
   ],
   "attention_bias": false,
   "attention_dropout": 0.0,
-  "bos_token_id": 1,
+  "bos_token_id": null,
   "draft_vocab_size": 32000,
-  "eos_token_id": 2,
+  "eagle_config": {
+    "eagle_aux_hidden_state_layer_ids": [
+      1,
+      23,
+      44
+    ],
+    "use_aux_hidden_state": true,
+    "use_input_layernorm_in_first_layer": true,
+    "use_last_layernorm": true,
+    "use_mtp_layernorm": false
+  },
+  "eos_token_id": null,
   "head_dim": 128,
   "hidden_act": "silu",
   "hidden_size": 5120,
   "initializer_range": 0.02,
   "intermediate_size": 32768,
-  "max_position_embeddings": 2048,
+  "max_position_embeddings": 1048576,
   "mlp_bias": false,
   "model_type": "llama",
   "num_attention_heads": 40,
   "num_hidden_layers": 1,
   "num_key_value_heads": 8,
-  "pad_token_id": 0,
   "pretraining_tp": 1,
   "rms_norm_eps": 1e-05,
   "rope_scaling": null,
-  "rope_theta": 10000.0,
+  "rope_theta": 500000.0,
   "tie_word_embeddings": false,
   "torch_dtype": "float32",
   "transformers_version": "4.53.2",

model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a49d698ac7f1ec7c8771eec4d16473674f16c8dc1c1089d9d41384c10d3d247b
+oid sha256:92f4188cdc145800df636a2c91f278f820fb56d25a8d3087b93cd55793f3b266
 size 4536860696

model-00002-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0a9df21baa435f3635340dd8686522e12398d45d9b5df1184e1c905611b08998
+oid sha256:4c070461f189c1e7e893f8f8c0d88e641b5282c8190262fc0c5fba2196ad0ad3
 size 2983281568