Fix example parameters
Browse files
README.md
CHANGED
|
@@ -194,8 +194,10 @@ grammar = LlamaGrammar.from_json_schema(json.dumps({
|
|
| 194 |
}
|
| 195 |
}))
|
| 196 |
|
| 197 |
-
llm = Llama(model_path="./Mistral-7B-Instruct-v0.3.IQ4_XS.gguf", n_gpu_layers=33, n_ctx=32768
|
| 198 |
response = llm.create_chat_completion(
|
|
|
|
|
|
|
| 199 |
messages = [
|
| 200 |
{
|
| 201 |
"role": "user",
|
|
@@ -228,6 +230,8 @@ response = llm.create_chat_completion(
|
|
| 228 |
print(json.loads(response["choices"][0]["text"]))
|
| 229 |
|
| 230 |
print(llm.create_chat_completion(
|
|
|
|
|
|
|
| 231 |
messages = [
|
| 232 |
{
|
| 233 |
"role": "user",
|
|
|
|
| 194 |
}
|
| 195 |
}))
|
| 196 |
|
| 197 |
+
llm = Llama(model_path="./Mistral-7B-Instruct-v0.3.IQ4_XS.gguf", n_gpu_layers=33, n_ctx=32768)
|
| 198 |
response = llm.create_chat_completion(
|
| 199 |
+
temperature = 0.0,
|
| 200 |
+
repeat_penalty = 1.1,
|
| 201 |
messages = [
|
| 202 |
{
|
| 203 |
"role": "user",
|
|
|
|
| 230 |
print(json.loads(response["choices"][0]["text"]))
|
| 231 |
|
| 232 |
print(llm.create_chat_completion(
|
| 233 |
+
temperature = 0.0,
|
| 234 |
+
repeat_penalty = 1.1,
|
| 235 |
messages = [
|
| 236 |
{
|
| 237 |
"role": "user",
|