nroggendorff
/

smallama

Text Generation

text-generation-inference

Inference Endpoints

Model card Files Files and versions Community

nroggendorff commited on 7 days ago

Commit

2f8fbcc

•

1 Parent(s): 44cdabd

Training loss: 21.4709

Files changed (2) hide show

config.json +5 -5
model.safetensors +2 -2

config.json CHANGED Viewed

@@ -9,15 +9,15 @@
   "eos_token_id": 2,
   "head_dim": 16,
   "hidden_act": "silu",
-  "hidden_size": 500,
   "initializer_range": 0.02,
-  "intermediate_size": 2000,
   "max_position_embeddings": 512,
   "mlp_bias": false,
   "model_type": "llama",
-  "num_attention_heads": 31,
-  "num_hidden_layers": 15,
-  "num_key_value_heads": 31,
   "pad_token_id": 1,
   "pretraining_tp": 1,
   "rms_norm_eps": 1e-05,

   "eos_token_id": 2,
   "head_dim": 16,
   "hidden_act": "silu",
+  "hidden_size": 1008,
   "initializer_range": 0.02,
+  "intermediate_size": 4032,
   "max_position_embeddings": 512,
   "mlp_bias": false,
   "model_type": "llama",
+  "num_attention_heads": 63,
+  "num_hidden_layers": 31,
+  "num_key_value_heads": 63,
   "pad_token_id": 1,
   "pretraining_tp": 1,
   "rms_norm_eps": 1e-05,

model.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f8fb82a021857a4ca78eb4ac644d7abb7d41198025d24f90eb78a5b91fc85c90
-size 447597344

 version https://git-lfs.github.com/spec/v1
+oid sha256:41392bcbd9ce5a0dc4aeb8cc39e132ef8046f41226a4f16bf740f2de4e532c62
+size 2435485272