Instructions to use MathMindsAGI/Test_context_pretrain with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MathMindsAGI/Test_context_pretrain with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="MathMindsAGI/Test_context_pretrain")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("MathMindsAGI/Test_context_pretrain")
model = AutoModelForCausalLM.from_pretrained("MathMindsAGI/Test_context_pretrain")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use MathMindsAGI/Test_context_pretrain with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MathMindsAGI/Test_context_pretrain"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MathMindsAGI/Test_context_pretrain",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/MathMindsAGI/Test_context_pretrain

SGLang

How to use MathMindsAGI/Test_context_pretrain 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 "MathMindsAGI/Test_context_pretrain" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MathMindsAGI/Test_context_pretrain",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "MathMindsAGI/Test_context_pretrain" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MathMindsAGI/Test_context_pretrain",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use MathMindsAGI/Test_context_pretrain with Docker Model Runner:
```
docker model run hf.co/MathMindsAGI/Test_context_pretrain
```

kmchiti commited on Mar 25

Commit

ef56d7d

verified ·

1 Parent(s): 180db9a

Upload artifacts/training/scripts/run_pretrain_id2-10_0.25easy_0.25medium_0.5hard.sh

Browse files

Files changed (1) hide show

artifacts/training/scripts/run_pretrain_id2-10_0.25easy_0.25medium_0.5hard.sh +62 -0

artifacts/training/scripts/run_pretrain_id2-10_0.25easy_0.25medium_0.5hard.sh ADDED Viewed

	@@ -0,0 +1,62 @@

+#!/usr/bin/env bash
+set -euo pipefail
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+REPO_ROOT="$(cd "${SCRIPT_DIR}/../../../.." && pwd)"
+cd "${REPO_ROOT}"
+# llamafactory-cli launches distributed training via `torchrun`, so the venv
+# bin dir must be on PATH even when the CLI itself is invoked by absolute path.
+export PATH="${REPO_ROOT}/.venv/bin:${PATH}"
+if [[ -n "${SLURM_CPUS_PER_TASK:-}" ]]; then
+    CPU_WORKERS_DEFAULT="${SLURM_CPUS_PER_TASK}"
+elif command -v nproc >/dev/null 2>&1; then
+    CPU_WORKERS_DEFAULT="$(nproc)"
+else
+    CPU_WORKERS_DEFAULT=1
+fi
+DEFAULT_PREPROCESSING_WORKERS="${PREPROCESSING_NUM_WORKERS:-${CPU_WORKERS_DEFAULT}}"
+DEFAULT_DATALOADER_WORKERS="${DATALOADER_NUM_WORKERS:-${CPU_WORKERS_DEFAULT}}"
+if [[ -z "${CUDA_VISIBLE_DEVICES:-}" ]]; then
+    echo "CUDA_VISIBLE_DEVICES must be set before running this script" >&2
+    exit 1
+fi
+LLAMA_BIN_DEFAULT="${REPO_ROOT}/.venv/bin/llamafactory-cli"
+DATASET_DIR_ROOT="${DATASET_DIR_ROOT:-data}"
+if [[ ! -x "${LLAMA_BIN_DEFAULT}" ]]; then
+    echo "Missing ${LLAMA_BIN_DEFAULT}. Run scripts/setup/install_local_llamafactory.sh first." >&2
+    exit 1
+fi
+if [[ ! -d "${REPO_ROOT}/${DATASET_DIR_ROOT}/composition/train" ]]; then
+    echo "Missing ${DATASET_DIR_ROOT}/composition/train. Run scripts/composition/prepare_hf_composition_data.sh first." >&2
+    exit 1
+fi
+if [[ ! -d "${REPO_ROOT}/${DATASET_DIR_ROOT}/composition/test" ]]; then
+    echo "Missing ${DATASET_DIR_ROOT}/composition/test. Run scripts/composition/prepare_hf_composition_data.sh first." >&2
+    exit 1
+fi
+export WANDB_PROJECT="${WANDB_PROJECT:-Interplay-LM-Reasoning}"
+export WANDB_ENTITY="${WANDB_ENTITY:-kmchiti}"
+DEFAULT_LLAMA_ARGS=(
+    "preprocessing_num_workers=${DEFAULT_PREPROCESSING_WORKERS}"
+    "dataloader_num_workers=${DEFAULT_DATALOADER_WORKERS}"
+)
+if [[ -n "${LLAMA_EXTRA_ARGS:-}" ]]; then
+    export LLAMA_EXTRA_ARGS="${DEFAULT_LLAMA_ARGS[*]} ${LLAMA_EXTRA_ARGS}"
+else
+    export LLAMA_EXTRA_ARGS="${DEFAULT_LLAMA_ARGS[*]}"
+fi
+EVAL_DATA_ROOT="${EVAL_DATA_ROOT:-${DATASET_DIR_ROOT}/composition/test}" \
+LLAMA_BIN="${LLAMA_BIN:-${LLAMA_BIN_DEFAULT}}" \
+LLAMA_CONFIG="scripts/composition/op-difficulty-10B/pt-diff2_10-tok10B-lr1e-4-bs512k-schedcos-minlr3e-5/id2-10_0.25easy_0.25medium_0.5hard.yaml" \
+    ./scripts/meta_run.sh --skip-rl "$@"