Text Generation
Transformers
Safetensors
GGUF
Korean
English
llama
3b
korean
from-scratch
orpo
instruction-tuned
preference-aligned
fp8
b200
Eval Results (legacy)
text-generation-inference
Instructions to use pathcosmos/frankenstallm with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use pathcosmos/frankenstallm with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="pathcosmos/frankenstallm")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("pathcosmos/frankenstallm") model = AutoModelForCausalLM.from_pretrained("pathcosmos/frankenstallm", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use pathcosmos/frankenstallm with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf pathcosmos/frankenstallm:Q4_K_M # Run inference directly in the terminal: llama cli -hf pathcosmos/frankenstallm:Q4_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf pathcosmos/frankenstallm:Q4_K_M # Run inference directly in the terminal: llama cli -hf pathcosmos/frankenstallm:Q4_K_M
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf pathcosmos/frankenstallm:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama cli -hf pathcosmos/frankenstallm:Q4_K_M
Use Docker
docker model run hf.co/pathcosmos/frankenstallm:Q4_K_M
- Llama new
- LM Studio
- Jan
- vLLM
How to use pathcosmos/frankenstallm with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "pathcosmos/frankenstallm" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "pathcosmos/frankenstallm", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/pathcosmos/frankenstallm:Q4_K_M
- SGLang
How to use pathcosmos/frankenstallm with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "pathcosmos/frankenstallm" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "pathcosmos/frankenstallm", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "pathcosmos/frankenstallm" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "pathcosmos/frankenstallm", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Ollama
How to use pathcosmos/frankenstallm with Ollama:
ollama run hf.co/pathcosmos/frankenstallm:Q4_K_M
- Unsloth Desktop
- Docker Model Runner
How to use pathcosmos/frankenstallm with Docker Model Runner:
docker model run hf.co/pathcosmos/frankenstallm:Q4_K_M
- Lemonade
How to use pathcosmos/frankenstallm with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull pathcosmos/frankenstallm:Q4_K_M
Run and chat with the model
lemonade run user.frankenstallm-Q4_K_M
List all available models
lemonade list
- Atomic Chat
File size: 6,065 Bytes
48ecd01 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 | #!/bin/bash
# ํ๊ตญ์ด ํ์ต ๋ฐ์ดํฐ ํํฉ ํ์ธ ์คํฌ๋ฆฝํธ
# ์ฉ๋: ํ๊ตญ์ด ๋ฐ์ดํฐ์
์ํ, ํ ํฌ๋์ด์ , ์๋ณธ ๋ฐ์ดํฐ ํ์ผ ํ์ธ
set -e
# ํ๋ก์ ํธ ๋ฃจํธ (์ด ์คํฌ๋ฆฝํธ ์คํ ์์น ๊ธฐ์ค)
PROJECT_ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
cd "${PROJECT_ROOT}"
echo "=== ํ๊ตญ์ด ํ์ต ๋ฐ์ดํฐ ํํฉ ==="
echo ""
# ============================================================================
# 1. ํ์ต์ฉ ๋ฐ์ด๋๋ฆฌ ๋ฐ์ดํฐ ํ์ธ
# ============================================================================
echo "[ ํ์ต ๋ฐ์ด๋๋ฆฌ ๋ฐ์ดํฐ ]"
check_binary_data() {
local file=$1
local name=$2
if [ -f "$file" ]; then
local size=$(du -h "$file" | cut -f1)
# Python + numpy memmap์ผ๋ก ํ ํฐ ์ ๊ณ์ฐ
# ๋ฐ์ด๋๋ฆฌ๋ uint32 ํํ๋ก ์ ์ฅ๋์ด ์์ (4 bytes per token)
local token_count=$(python3 -c "
import numpy as np
try:
data = np.memmap('$file', dtype=np.uint32, mode='r')
print(len(data))
except Exception as e:
print('error')
" 2>/dev/null || echo "error")
if [ "$token_count" != "error" ] && [ ! -z "$token_count" ]; then
# ํ ํฐ ์๋ฅผ ํฌ๋งทํ
(์ฒ ๋จ์ ์ผํ)
local formatted_tokens=$(printf "%'d" "$token_count")
# 1B ๋ชจ๋ธ ํ์ต ์คํ
๊ณ์ฐ
# tokens_per_step = batch_size * grad_accum * seq_len * num_gpus
# = 8 * 4 * 4096 * 8 = 1,048,576 tokens/step
local tokens_per_step=1048576
local estimated_steps=$((token_count / tokens_per_step))
printf " %-20s : ์กด์ฌ (%s, %'d ํ ํฐ, ~%'d steps)\n" \
"$name" "$size" "$token_count" "$estimated_steps"
else
printf " %-20s : ์กด์ฌ (%s, ํ ํฐ ๊ณ์ฐ ์คํจ)\n" "$name" "$size"
fi
else
printf " %-20s : ์์\n" "$name"
fi
}
check_binary_data "data/korean_train.bin" "korean_train.bin"
check_binary_data "data/korean_val.bin" "korean_val.bin"
check_binary_data "data/train.bin" "train.bin"
check_binary_data "data/val.bin" "val.bin"
echo ""
# ============================================================================
# 2. ํ ํฌ๋์ด์ ํ์ธ
# ============================================================================
echo "[ ํ ํฌ๋์ด์ ]"
check_tokenizer() {
local dir=$1
local name=$2
if [ -d "$dir" ]; then
local files=$(find "$dir" -type f | wc -l)
printf " %-20s : ์กด์ฌ (%d๊ฐ ํ์ผ)\n" "$name" "$files"
else
printf " %-20s : ์์\n" "$name"
fi
}
check_tokenizer "tokenizer/korean_sp" "korean_sp"
check_tokenizer "tokenizer" "default tokenizer"
echo ""
# ============================================================================
# 3. ์๋ณธ ๋ฐ์ดํฐ ๋๋ ํ ๋ฆฌ ํ์ธ
# ============================================================================
echo "[ ์๋ณธ ๋ฐ์ดํฐ ]"
check_raw_data() {
local dir=$1
local name=$2
if [ -d "$dir" ]; then
local file_count=$(find "$dir" -maxdepth 1 -type f | wc -l)
local total_size=$(du -sh "$dir" 2>/dev/null | cut -f1)
if [ $file_count -eq 0 ]; then
printf " %-20s : ์์ (๋๋ ํ ๋ฆฌ๋ง ์กด์ฌ, 0 ํ์ผ)\n" "$name"
else
printf " %-20s : %'d ํ์ผ (%s)\n" "$name" "$file_count" "$total_size"
fi
else
printf " %-20s : ์์\n" "$name"
fi
}
check_raw_data "data/raw/cc100_ko" "cc100_ko/"
check_raw_data "data/raw/c4_ko" "c4_ko/"
check_raw_data "data/raw/namuwiki_ko" "namuwiki_ko/"
# ์ํค ๋ฐ์ดํฐ๋ raw/ ์ง์ ํ์
echo ""
echo "[ ์ํคํผ๋์ ๋ฐ์ดํฐ ]"
ko_wiki_count=$(find "data/raw" -maxdepth 1 -name "ko_wiki_*.txt" | wc -l)
en_wiki_count=$(find "data/raw" -maxdepth 1 -name "en_wiki_*.txt" | wc -l)
ko_wiki_size=$(du -sh "data/raw" 2>/dev/null | cut -f1)
if [ $ko_wiki_count -gt 0 ]; then
printf " %-20s : %'d ํ์ผ\n" "ko_wiki" "$ko_wiki_count"
fi
if [ $en_wiki_count -gt 0 ]; then
printf " %-20s : %'d ํ์ผ\n" "en_wiki" "$en_wiki_count"
fi
echo ""
# ============================================================================
# 4. ์ข
ํฉ ์ํ ์์ฝ
# ============================================================================
echo "[ ์ข
ํฉ ์ํ ]"
# ํ์ต์ฉ ๋ฐ์ด๋๋ฆฌ ๋ฐ์ดํฐ ํ์ธ
binary_ready=false
if [ -f "data/korean_train.bin" ] && [ -f "data/korean_val.bin" ]; then
binary_ready=true
elif [ -f "data/train.bin" ] && [ -f "data/val.bin" ]; then
binary_ready=true
fi
# ํ ํฌ๋์ด์ ํ์ธ
tokenizer_ready=false
if [ -d "tokenizer/korean_sp" ] && [ -f "tokenizer/korean_sp/tokenizer.model" ]; then
tokenizer_ready=true
fi
# ์๋ณธ ๋ฐ์ดํฐ ํ์ธ
raw_ready=false
if [ -d "data/raw/c4_ko" ] || [ -d "data/raw/namuwiki_ko" ] || [ -d "data/raw/cc100_ko" ]; then
count=$(find "data/raw/c4_ko" -maxdepth 1 -type f 2>/dev/null | wc -l)
count=$((count + $(find "data/raw/namuwiki_ko" -maxdepth 1 -type f 2>/dev/null | wc -l)))
count=$((count + $(find "data/raw/cc100_ko" -maxdepth 1 -type f 2>/dev/null | wc -l)))
if [ $count -gt 0 ]; then
raw_ready=true
fi
fi
printf " ํ์ต์ฉ ๋ฐ์ด๋๋ฆฌ : %s\n" "$([ "$binary_ready" = true ] && echo "โ ์ค๋น๋จ" || echo "โ ๋ฏธ์ค๋น")"
printf " ํ ํฌ๋์ด์ : %s\n" "$([ "$tokenizer_ready" = true ] && echo "โ ์ค๋น๋จ" || echo "โ ๋ฏธ์ค๋น")"
printf " ์๋ณธ ๋ฐ์ดํฐ : %s\n" "$([ "$raw_ready" = true ] && echo "โ ์ค๋น๋จ" || echo "โ ๋ฏธ์ค๋น")"
echo ""
# ============================================================================
# 5. ํ์ต ์ค์ ํ๋ผ๋ฏธํฐ ์ ๋ณด
# ============================================================================
echo "[ ํ์ต ์ค์ (1B ๋ชจ๋ธ ๊ธฐ์ค) ]"
echo " ๋ฐฐ์น ์ฌ์ด์ฆ : 8"
echo " ์ํ์ค ๊ธธ์ด : 4096"
echo " GPU ์ : 8"
echo " ๊ทธ๋๋์ธํธ ๋์ : 4"
echo " ํ ํฐ/์คํ
: 8 ร 4 ร 4096 ร 8 = 1,048,576"
echo ""
echo "=== ๊ฒ์ฌ ์๋ฃ ==="
|