Shouldn't CodeLlama 34B have 16K context and rope_theta 1M? (#3)

- Shouldn't CodeLlama 34B have 16K context and rope_theta 1M? (1a41a8c0a56a9894057366eec405db87443a9fa0)


Co-authored-by: Tom Jobbins <TheBloke@users.noreply.huggingface.co>

config.json

@@ -9,7 +9,7 @@
   "hidden_size": 8192,
   "initializer_range": 0.02,
   "intermediate_size": 22016,
-  "max_position_embeddings": 2048,
+  "max_position_embeddings": 16384,
   "model_type": "llama",
   "num_attention_heads": 64,
   "num_hidden_layers": 48,
@@ -18,6 +18,7 @@
   "pretraining_tp": 1,
   "rms_norm_eps": 1e-05,
   "rope_scaling": null,
+  "rope_theta": 1000000,
   "tie_word_embeddings": false,
   "torch_dtype": "float16",
   "transformers_version": "4.31.0",