qpwo

def universal_summary(x, depth=0, max_depth=4, max_items=5, max_chars=350):
    """
    Provides a concise summary of any Python object.
    
    Args:
        x: Any Python object
        depth: Current recursion depth (internal use)
        max_depth: Maximum recursion depth
        max_items: Maximum number of items to show for collections
        max_chars: Maximum characters in the final output

qpwo

import torch
from torch.utils.cpp_extension import load_inline

# Define C++ source code
cpp_source = """
#include <torch/extension.h>

torch::Tensor add_one(torch::Tensor input) {
    return input + 1;
}

qpwo

• EfficientZero V2: A general sample-efficient RL framework excels in diverse control tasks (discrete/continuous, visual/low-dimensional) outperforming SoTA, including DreamerV3, in 50/66 benchmarks.
• Gambling-Based Confidence Sequences: A novel gambling framework constructs tight, non-asymptotic confidence sequences for bounded random vectors, including categorical and probability-vector-valued observations, outperforming existing methods like the posterior-prior ratio martingale.
• In-Context Learning Circuits: Mechanistic study reveals how induction heads, key to in-context learning in transformers, emerge through interactions of three identified sub-circuits during training.
• Explaining Probabilistic Models with Distributional Values: This paper introduces distributional values, generalising cooperative game theory and value operators to provide fine-grained explanations of probabilistic models like vision and language models by tracking changes in model output rather than s

qpwo

/**
 * my-validator.test.ts
 *
 * Test suite for my-validator.ts library
 */

import { objSchema, arrSchema, unionSchema, validate, Infer } from './my-validator'

// Test 1: Basic object schema validation
function testBasicObject() {

qpwo

import mpmath as mp
import math

# Set high precision
mp.mp.dps = 50

def calculate_collision_probability(n, d):
    """
    Calculate probability of at least one collision when choosing n items from d possibilities
    Using the birthday paradox/collision probability formula: 1 - exp(-n(n-1)/(2d))

qpwo

/** Run different types of jobs via postgres queue.
 *  Aims for exactly-once completion of (job, key) pair.
 */
import 'dotenv/config'
import { Pool, PoolClient } from 'pg'
import { JobStatus } from 'shared'

type StatusObj<TData = any, TResult = any> = {
  jobId: string
  jobType: string

qpwo

curl https://rclone.org/install.sh | sudo bash
mkdir ~/.config/rclone
echo "
[r2]
type = s3
provider = Cloudflare
access_key_id = $a
secret_access_key = $b
region = auto
endpoint = https://$c.r2.cloudflarestorage.com

qpwo

Why does this run instantly after first trial?

import subprocess
import time
import shutil
import os


def run_rclone_command(command, num_trials=3, dest_path_base="/home/ubuntu/hff/405b-base"):
    """

qpwo

# modal run issue.py
import os
import subprocess
import time

import modal
import modal.experimental
import modal.gpu
import modal.volume

qpwo

pip install huggingface_hub[hf_transfer]
export HF_HUB_ENABLE_HF_TRANSFER=1
model_name=meta-llama/Llama-3.1-405B
localdir=$(realpath ~/hff/405b)
huggingface-cli download --max-workers=8 --include="model-???[02468][02468]-of-?????.safetensors" --local-dir=$localdir $model_name & sleep 6
huggingface-cli download --max-workers=8 --include="model-???[02468][13579]-of-?????.safetensors" --local-dir=$localdir $model_name & sleep 6
huggingface-cli download --max-workers=8 --include="model-???[13579][02468]-of-?????.safetensors" --local-dir=$localdir $model_name & sleep 6
huggingface-cli download --max-workers=8 --include="model-???[13579][13579]-of-?????.safetensors"  --local-dir=$localdir $model_name & sleep 0