Upload PhiForCausalLM

config.json

@@ -1,29 +1,36 @@
 {
+  "activation_function": "gelu_new",
   "architectures": [
     "PhiForCausalLM"
   ],
   "attention_dropout": 0.0,
-  "bos_token_id": null,
+  "attn_pdrop": 0.0,
   "embd_pdrop": 0.0,
-  "eos_token_id": null,
+  "flash_attn": false,
+  "flash_rotary": false,
+  "fused_dense": false,
   "hidden_act": "gelu_new",
-  "hidden_size": 2048,
   "initializer_range": 0.02,
   "intermediate_size": 8192,
   "layer_norm_eps": 1e-05,
-  "max_position_embeddings": 2048,
+  "layer_norm_epsilon": 1e-05,
   "model_type": "phi",
+  "n_embd": 1024,
+  "n_head": 16,
+  "n_head_kv": null,
+  "n_inner": null,
   "n_layer": 1,
-  "num_attention_heads": 32,
-  "num_hidden_layers": 24,
-  "num_key_value_heads": 32,
+  "n_positions": 2048,
+  "num_key_value_heads": null,
+  "pad_vocab_size_multiple": 64,
   "partial_rotary_factor": 0.5,
   "qk_layernorm": false,
   "resid_pdrop": 0.0,
   "rope_scaling": null,
   "rope_theta": 10000.0,
+  "rotary_dim": 32,
   "tie_word_embeddings": false,
-  "torch_dtype": "bfloat16",
+  "torch_dtype": "float32",
   "transformers_version": "4.50.0",
   "use_cache": true,
   "vocab_size": 51200

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e4718cf05cf16ac9b88835b963eca12b713f89542a652f3998760c32b533db0f
-size 2836579040
+oid sha256:7879646e2b1e0e2b1ccfe9d67dfdfb3ed82c53fb452f54b7a7bcc2b301c09258
+size 469800176