feat: add tranning in start container

constants.py

@@ -37,15 +37,13 @@ MAX_NEW_TOKENS = CONTEXT_WINDOW_SIZE  # int(CONTEXT_WINDOW_SIZE/4)
 
 #### If you get a "not enough space in the buffer" error, you should reduce the values below, start with half of the original values and keep halving the value until the error stops appearing
 
-N_GPU_LAYERS = 83  # Llama-2-70B has 83 layers
+N_GPU_LAYERS = 100  # Llama-2-70B has 83 layers
 N_BATCH = CONTEXT_WINDOW_SIZE
 
 ### From experimenting with the Llama-2-7B-Chat-GGML model on 8GB VRAM, these values work:
 # N_GPU_LAYERS = 20
 # N_BATCH = 512
 
-
-
 # https://python.langchain.com/en/latest/_modules/langchain/document_loaders/excel.html#UnstructuredExcelLoader
 DOCUMENT_MAP = {
     ".txt": TextLoader,
@@ -100,11 +98,11 @@ EMBEDDING_MODEL_NAME = "hkunlp/instructor-large"  # Uses 1.5 GB of VRAM (High Ac
 #### (FOR GGUF MODELS)
 ####
 
-# MODEL_ID = "TheBloke/Llama-2-13b-Chat-GGUF"
-# MODEL_BASENAME = "llama-2-13b-chat.Q4_K_M.gguf"
+MODEL_ID = "TheBloke/Llama-2-13b-Chat-GGUF"
+MODEL_BASENAME = "llama-2-13b-chat.Q4_K_M.gguf"
 
-MODEL_ID = "TheBloke/Llama-2-7b-Chat-GGUF"
-MODEL_BASENAME = "llama-2-7b-chat.Q4_K_M.gguf"
+# MODEL_ID = "TheBloke/Llama-2-7b-Chat-GGUF"
+# MODEL_BASENAME = "llama-2-7b-chat.Q4_K_M.gguf"
 
 # MODEL_ID = "TheBloke/Mistral-7B-Instruct-v0.1-GGUF"
 # MODEL_BASENAME = "mistral-7b-instruct-v0.1.Q8_0.gguf"