| """ | |
| The most atomic way to train and inference a GPT in pure, dependency-free Python. | |
| This file is the complete algorithm. | |
| Everything else is just efficiency. | |
| @karpathy | |
| """ | |
| import os # os.path.exists | |
| import math # math.log, math.exp |
| import MySharedWorker from './mySharedWorker.ts?sharedworker'; // using Vite shared worker import | |
| import { messagePortInterface } from './messagePortInterface'; | |
| // | |
| // We need to define the interfaces that we use on both sides of the message port. | |
| // | |
| // This is the interface we use for the frontend, i.e. when the worker wants to call | |
| // a function remotely against the UI thread. | |
| export type MainThreadInterface = { |
| GET /beta/deviceLocalCredentials/[DEVICE-ID]?$select=credentials HTTP/1.1 | |
| ocp-client-version: 1.0 | |
| client-request-id: 96cbfa59-dbfc-4a92-b261-7f77bd8f4b9b | |
| ocp-client-name: Get-LapsAADPassword Windows LAPS Cmdlet | |
| User-Agent: Mozilla/5.0 (Windows NT 10.0; Microsoft Windows 10.0.22621; en-US) PowerShell/5.1.22621.963 Invoke-MgGraphRequest | |
| SdkVersion: graph-powershell/1.26.0, Graph-dotnet-1.25.1 | |
| FeatureFlag: 00000047 | |
| Cache-Control: no-store, no-cache | |
| Authorization: Bearer [AAD-JWT-HERE] | |
| Accept-Encoding: gzip |
| mp=1; size=7B; # to run 7B | |
| mp=8; size=65B; # to run 65B | |
| for seed in $(randint 1000000) | |
| do | |
| export TARGET_FOLDER=~/ml/data/llama/LLaMA | |
| time python3 -m torch.distributed.run --nproc_per_node $mp example.py --ckpt_dir $TARGET_FOLDER/$size --tokenizer_path $TARGET_FOLDER/tokenizer.model --seed $seed --max_seq_len 2048 --max_gen_len 2048 --count 0 | tee -a ${size}_startrek.txt | |
| done |
| { | |
| "method": "ProtocolsConfigure", | |
| "protocol": "https://identity.foundation/schemas/wallet", | |
| "protocolVersion": "0.0.1", | |
| "protocolDefinition": { | |
| "labels": { | |
| "persona": { | |
| "schema": "https://schema.org/Person", | |
| "dataFormat": [ | |
| "application/json" |
| { | |
| "interface": "Protocols", | |
| "method": "Configure", | |
| "definition": { | |
| "protocol": "https://chat.protocol/", | |
| "types": { | |
| "thread": { | |
| "schema": "https://chat.protocol/schemas/thread", | |
| "dataFormat": [ | |
| "application/json" |
| { | |
| "method": "ProtocolsConfigure", | |
| "protocol": "https://decentralized-music.org/protocol", | |
| "protocolVersion": "1.0.0", | |
| "protocolDefinition": { | |
| "labels": { | |
| "playlist": { | |
| "schema": "https://decentralized-music.org/protocol/playlist", | |
| "dataFormat": [ | |
| "application/json" |
| ---- MODULE Filters ---- | |
| EXTENDS TLC, Integers | |
| VARIABLE push_msgs, emails, i, filtered, pushed | |
| vars == <<push_msgs, emails, i, filtered, pushed>> | |
| set ++ x == set \union {x} | |
| set -- x == set \ {x} | |
| TypeInv == | |
| /\ emails \subseteq (1..3) |
I have an updated version of this on my blog here: https://chrisamico.com/blog/2023-01-14/python-setup/.
This is my recommended Python setup, as of Fall 2022. The Python landscape can be a confusing mess of overlapping tools that sometimes don't work well together. This is an effort to standardize our approach and environments.
