Text Generation
Transformers
Safetensors
olmo3
code
reasoning
lora-merged
livecodebench
conversational
Instructions to use modrill/code-think-o7b-20260908 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use modrill/code-think-o7b-20260908 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="modrill/code-think-o7b-20260908") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("modrill/code-think-o7b-20260908") model = AutoModelForCausalLM.from_pretrained("modrill/code-think-o7b-20260908", device_map="auto") 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 Settings
- vLLM
How to use modrill/code-think-o7b-20260908 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "modrill/code-think-o7b-20260908" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "modrill/code-think-o7b-20260908", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/modrill/code-think-o7b-20260908
- SGLang
How to use modrill/code-think-o7b-20260908 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 "modrill/code-think-o7b-20260908" \ --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": "modrill/code-think-o7b-20260908", "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 "modrill/code-think-o7b-20260908" \ --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": "modrill/code-think-o7b-20260908", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use modrill/code-think-o7b-20260908 with Docker Model Runner:
docker model run hf.co/modrill/code-think-o7b-20260908
Add provenance/RUN_IDENTITY.json
Browse files- provenance/RUN_IDENTITY.json +108 -0
provenance/RUN_IDENTITY.json
ADDED
|
@@ -0,0 +1,108 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"schema": "RUN_IDENTITY_V1",
|
| 3 |
+
"run_id": "t30b2507-o7b-think-v4-tail151643",
|
| 4 |
+
"corrects": "PAIRED_V4",
|
| 5 |
+
"model": {
|
| 6 |
+
"hf_id": "allenai/Olmo-3-1025-7B",
|
| 7 |
+
"revision": "a81bae42db3975be1671e27b9c9a56da1a9f980f",
|
| 8 |
+
"local_path": "/workspace/code-sft-infra/models/olmo-3-1025-7b",
|
| 9 |
+
"tied_embeddings": false
|
| 10 |
+
},
|
| 11 |
+
"data": {
|
| 12 |
+
"path": "/workspace/code-sft-runs/t30b2507-q4b-tn-v2-lr1e4-e1-20260907/data/THINK_PAIRED_V4_2EP_O7B.parquet",
|
| 13 |
+
"rows": 9430,
|
| 14 |
+
"sha256": "17554a69f4b99d6d0b0861bece9e959a664084e00dc38ee76a5ba711e2d8da60",
|
| 15 |
+
"census": "/workspace/code-sft-runs/t30b2507-q4b-tn-v2-lr1e4-e1-20260907/data/TOKEN_CENSUS_V4_O7B_THINK.json",
|
| 16 |
+
"physical_2ep": "concat(selected_v4_table, selected_v4_table) same order; token columns overlaid from TOKEN_CENSUS_V4; O7B also rewrites renderer_id",
|
| 17 |
+
"source_1ep_path": "/workspace/code-sft-runs/t30b2507-q4b-tn-v2-lr1e4-e1-20260907/data/THINK_PAIRED_V4.parquet",
|
| 18 |
+
"source_1ep_sha256": "cf37c85413f69339508004611e196e42abbc2bddbd3462c209a1b97e1d820002",
|
| 19 |
+
"source_1ep_rows": 4715
|
| 20 |
+
},
|
| 21 |
+
"seeds": {
|
| 22 |
+
"train": 42,
|
| 23 |
+
"eval": 3407
|
| 24 |
+
},
|
| 25 |
+
"parameters": {
|
| 26 |
+
"epochs": 2,
|
| 27 |
+
"target_assistant_tokens": 63378772,
|
| 28 |
+
"active_tokens_1ep": 31689386,
|
| 29 |
+
"checkpoint_target_tokens": [
|
| 30 |
+
31722729,
|
| 31 |
+
47449622,
|
| 32 |
+
63378772
|
| 33 |
+
],
|
| 34 |
+
"milestone_updates": [
|
| 35 |
+
484,
|
| 36 |
+
724,
|
| 37 |
+
968
|
| 38 |
+
],
|
| 39 |
+
"warmup_target_tokens": 3802726,
|
| 40 |
+
"learning_rate": 0.0001,
|
| 41 |
+
"row_lr": 0.0001,
|
| 42 |
+
"optimizer": "adamw",
|
| 43 |
+
"optimizer_betas": [
|
| 44 |
+
0.9,
|
| 45 |
+
0.95
|
| 46 |
+
],
|
| 47 |
+
"optimizer_eps": 1e-08,
|
| 48 |
+
"lora_rank": 64,
|
| 49 |
+
"lora_alpha": 128,
|
| 50 |
+
"lora_dropout": 0.0,
|
| 51 |
+
"lora_target_modules": [
|
| 52 |
+
"q_proj",
|
| 53 |
+
"k_proj",
|
| 54 |
+
"v_proj",
|
| 55 |
+
"o_proj",
|
| 56 |
+
"gate_proj",
|
| 57 |
+
"up_proj",
|
| 58 |
+
"down_proj"
|
| 59 |
+
],
|
| 60 |
+
"lora_weight_decay": 0.1,
|
| 61 |
+
"row_weight_decay": 0.0,
|
| 62 |
+
"max_grad_norm": 1.0,
|
| 63 |
+
"tokens_per_optimizer_update": 65536,
|
| 64 |
+
"packing": false,
|
| 65 |
+
"truncation": false,
|
| 66 |
+
"repeat_examples": false,
|
| 67 |
+
"precision": "bfloat16",
|
| 68 |
+
"trainable_dtype": "float32",
|
| 69 |
+
"gradient_checkpointing": true,
|
| 70 |
+
"eos_weight": 1.0,
|
| 71 |
+
"scheduler": "cosine_by_assistant_target_token_dose",
|
| 72 |
+
"worker_lr_scale": "rstar_dual_worker_v4._lr_scale",
|
| 73 |
+
"supervised_tail": [
|
| 74 |
+
100257
|
| 75 |
+
],
|
| 76 |
+
"trainable_token_indices": {
|
| 77 |
+
"embed_tokens": [
|
| 78 |
+
100257
|
| 79 |
+
],
|
| 80 |
+
"lm_head": [
|
| 81 |
+
100257
|
| 82 |
+
]
|
| 83 |
+
},
|
| 84 |
+
"untie_tied_embeddings": true,
|
| 85 |
+
"context": 32768,
|
| 86 |
+
"eval": {
|
| 87 |
+
"suite": "DEV256",
|
| 88 |
+
"seed": 3407,
|
| 89 |
+
"temperature": 0.6,
|
| 90 |
+
"top_p": 0.95,
|
| 91 |
+
"top_k": 20,
|
| 92 |
+
"ctx": 32768,
|
| 93 |
+
"stop_ids": [
|
| 94 |
+
100257,
|
| 95 |
+
100265
|
| 96 |
+
],
|
| 97 |
+
"base_reeval_same_mode": true
|
| 98 |
+
}
|
| 99 |
+
},
|
| 100 |
+
"code": {
|
| 101 |
+
"path": "/workspace/tools/rstar_dual_worker_v4.py",
|
| 102 |
+
"commit_or_sha256": "97a298e9ee4f0bb571baa88c7d75285ade2ea2a8a9b5875b35db25b5db9b3842",
|
| 103 |
+
"template": "/workspace/code-sft-runs/t30b2507-q4b-tn-v2-lr1e4-e1-20260907/templates/O7B-THINK.template_v3.json"
|
| 104 |
+
},
|
| 105 |
+
"started_at": "2026-09-08T07:50:11Z",
|
| 106 |
+
"endpoint": "2EP_ONLY @63378772 assistant tokens; no checkpoint picking",
|
| 107 |
+
"status": "RUNNING"
|
| 108 |
+
}
|