Good question! I am collecting human data on how quantization affects outputs. See here for more information: ggerganov/llama.cpp#5962
In the meantime, use the largest that fully fits in your GPU. If you can comfortably fit Q4_K_S, try using a model with more parameters.
See the wiki upstream: https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix
Code is clean if it can be understood easily – by everyone on the team. Clean code can be read and enhanced by a developer other than its original author. With understandability comes readability, changeability, extensibility and maintainability.
| import autogen | |
| from user_proxy_webagent import UserProxyWebAgent | |
| import asyncio | |
| config_list = [ | |
| { | |
| "model": "gpt-3.5-turbo", | |
| # "api_key": "<YOUR KEY HERE>" | |
| } | |
| ] |
People
 :bowtie: |
😄 :smile: |
😆 :laughing: |
|---|---|---|
😊 :blush: |
😃 :smiley: |
:relaxed: |
😏 :smirk: |
😍 :heart_eyes: |
😘 :kissing_heart: |
😚 :kissing_closed_eyes: |
😳 :flushed: |
😌 :relieved: |
😆 :satisfied: |
😁 :grin: |
😉 :wink: |
😜 :stuck_out_tongue_winking_eye: |
😝 :stuck_out_tongue_closed_eyes: |
😀 :grinning: |
😗 :kissing: |
😙 :kissing_smiling_eyes: |
😛 :stuck_out_tongue: |
| def is_safe(graph, colours, colour, current): | |
| neighbours = graph[current] | |
| for neighbour in neighbours: | |
| if colour == colours[neighbour - 1]: | |
| return False | |
| return True | |
| def graph_traversal(graph, colours, vertices, current): | |
| if current == vertices + 1: | |
| return True |
| import heapq | |
| from typing import List | |
| def djikstra(graph_inner, starting, vertices) -> List: | |
| pq = [] | |
| dist = [float('inf') for i in range(vertices)] | |
| dist[starting] = 0 | |
| print(dist) | |
| heapq.heappush(pq, [0, starting]) |
| class Node: | |
| def __init__(self, data: list, level, f_val): | |
| self.data = data | |
| self.level = level | |
| self.f_val = f_val | |
| def generate_children(self): | |
| children = [] | |
| values = [[0, 1], [1, 0], [-1, 0], [0, -1]] | |
| row, col = self.find() |
Source: https://medium.com/@Mahmoud_Zalt/eloquent-relationships-cheat-sheet-5155498c209
| One to one ( 1-1) |
One to many ( 1-n) |
Poly one to many ( 1x-n) |
Many to many ( n-n) |
Poly many to many ( nx-n) |
|
|---|---|---|---|---|---|
| Number of models | 2 only | 2 only | 3 and above | 2 only | 3 and above |
| Number of tables | 2 (1/model) | 2 (1/model) | 3+ (1/model) | 3 (1/model + pivot) | 4+ (1/model + pivot) |
| Pivot table | - | - | - | required |
