eas/
openchat:latest

89 1 year ago

q4_k_m quantization only. Now using 8k context size

... /
params
624d5605c62b · 190B
{
"num_ctx": 8192,
"stop": [
"<|end_of_turn|>",
"<|end\\_of\\_turn|>",
"<|end_of\\_turn|>",
"GPT4User",
"User 1:",
"GPT4 User",
"Reddit User",
"GPT4:",
"GPT4 Correct User"
]
}