Text Generation
Transformers
Safetensors
Chinese
English
samai_27b
pondernet
mixture-of-experts
adaptive-computation
multi-token-prediction
qwen3.8
conversational
custom_code
Instructions to use tchbcb/samai-27b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use tchbcb/samai-27b with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="tchbcb/samai-27b", trust_remote_code=True) messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# pip install -U transformers accelerate # Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("tchbcb/samai-27b", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use tchbcb/samai-27b with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "tchbcb/samai-27b" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "tchbcb/samai-27b", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/tchbcb/samai-27b
- SGLang
How to use tchbcb/samai-27b 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 "tchbcb/samai-27b" \ --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": "tchbcb/samai-27b", "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 "tchbcb/samai-27b" \ --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": "tchbcb/samai-27b", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use tchbcb/samai-27b with Docker Model Runner:
docker model run hf.co/tchbcb/samai-27b
backup s27_train_metrics.json
Browse files
sft-adapter/s27_train_metrics.json
ADDED
|
@@ -0,0 +1,258 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"train_loss": 0.5060088133367782,
|
| 3 |
+
"train_runtime": 4619.2889,
|
| 4 |
+
"trainable_params": 124657736,
|
| 5 |
+
"feats": 831,
|
| 6 |
+
"epochs": 1,
|
| 7 |
+
"mtp_loss_weight": 0.0,
|
| 8 |
+
"mtp_note": "MTP loss disabled for T4 VRAM (full-seq fp32 logits chain); MTP is bypassed at inference & dropped in GGUF",
|
| 9 |
+
"maxlen": 256,
|
| 10 |
+
"max_ponder_steps": 4,
|
| 11 |
+
"eos_id": 248046,
|
| 12 |
+
"device_map": {
|
| 13 |
+
"model.language_model.embed_tokens": "0",
|
| 14 |
+
"model.language_model.layers.0": "0",
|
| 15 |
+
"model.language_model.layers.1": "0",
|
| 16 |
+
"model.language_model.layers.2": "0",
|
| 17 |
+
"model.language_model.layers.3": "0",
|
| 18 |
+
"model.language_model.layers.4": "0",
|
| 19 |
+
"model.language_model.layers.5": "0",
|
| 20 |
+
"model.language_model.layers.6": "0",
|
| 21 |
+
"model.language_model.layers.7": "0",
|
| 22 |
+
"model.language_model.layers.8": "0",
|
| 23 |
+
"model.language_model.layers.9": "0",
|
| 24 |
+
"model.language_model.layers.10": "0",
|
| 25 |
+
"model.language_model.layers.11": "0",
|
| 26 |
+
"model.language_model.layers.12": "0",
|
| 27 |
+
"model.language_model.layers.13": "0",
|
| 28 |
+
"model.language_model.layers.14": "0",
|
| 29 |
+
"model.language_model.layers.15": "0",
|
| 30 |
+
"model.language_model.layers.16": "0",
|
| 31 |
+
"model.language_model.layers.17": "0",
|
| 32 |
+
"model.language_model.layers.18": "0",
|
| 33 |
+
"model.language_model.layers.19": "0",
|
| 34 |
+
"model.language_model.layers.20": "0",
|
| 35 |
+
"model.language_model.layers.21": "0",
|
| 36 |
+
"model.language_model.layers.22": "0",
|
| 37 |
+
"model.language_model.layers.23": "0",
|
| 38 |
+
"model.language_model.layers.24": "0",
|
| 39 |
+
"model.language_model.layers.25": "0",
|
| 40 |
+
"model.language_model.layers.26": "0",
|
| 41 |
+
"model.language_model.layers.27": "0",
|
| 42 |
+
"model.language_model.layers.28": "0",
|
| 43 |
+
"model.language_model.layers.29": "0",
|
| 44 |
+
"model.language_model.layers.30": "0",
|
| 45 |
+
"model.language_model.layers.31": "0",
|
| 46 |
+
"model.language_model.layers.32": "0",
|
| 47 |
+
"model.language_model.layers.33": "0",
|
| 48 |
+
"model.language_model.layers.34": "0",
|
| 49 |
+
"model.language_model.layers.35": "0",
|
| 50 |
+
"model.language_model.layers.36": "0",
|
| 51 |
+
"model.language_model.layers.37": "0",
|
| 52 |
+
"model.language_model.layers.38": "1",
|
| 53 |
+
"model.language_model.layers.39": "1",
|
| 54 |
+
"model.language_model.layers.40": "1",
|
| 55 |
+
"model.language_model.layers.41": "1",
|
| 56 |
+
"model.language_model.layers.42": "1",
|
| 57 |
+
"model.language_model.layers.43": "1",
|
| 58 |
+
"model.language_model.layers.44": "1",
|
| 59 |
+
"model.language_model.layers.45": "1",
|
| 60 |
+
"model.language_model.layers.46": "1",
|
| 61 |
+
"model.language_model.layers.47": "1",
|
| 62 |
+
"model.language_model.layers.48": "1",
|
| 63 |
+
"model.language_model.layers.49": "1",
|
| 64 |
+
"model.language_model.layers.50": "1",
|
| 65 |
+
"model.language_model.layers.51": "1",
|
| 66 |
+
"model.language_model.layers.52": "1",
|
| 67 |
+
"model.language_model.layers.53": "1",
|
| 68 |
+
"model.language_model.layers.54": "1",
|
| 69 |
+
"model.language_model.layers.55": "1",
|
| 70 |
+
"model.language_model.layers.56": "1",
|
| 71 |
+
"model.language_model.layers.57": "1",
|
| 72 |
+
"model.language_model.layers.58": "1",
|
| 73 |
+
"model.language_model.layers.59": "1",
|
| 74 |
+
"model.language_model.layers.60": "1",
|
| 75 |
+
"model.language_model.layers.61": "1",
|
| 76 |
+
"model.language_model.layers.62": "1",
|
| 77 |
+
"model.language_model.layers.63": "1",
|
| 78 |
+
"model.language_model.norm": "1",
|
| 79 |
+
"mtp": "1",
|
| 80 |
+
"lm_head": "1",
|
| 81 |
+
"model.visual": "cpu",
|
| 82 |
+
"model.language_model.rotary_emb": "0"
|
| 83 |
+
},
|
| 84 |
+
"ponder_stats": {
|
| 85 |
+
"layers": [
|
| 86 |
+
{
|
| 87 |
+
"layer": 35,
|
| 88 |
+
"executed": 2,
|
| 89 |
+
"steps_mean": 1.5032258033752441,
|
| 90 |
+
"moe_active": 2.9612903594970703,
|
| 91 |
+
"moe_aux": 3.4045186042785645
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"layer": 39,
|
| 95 |
+
"executed": 4,
|
| 96 |
+
"steps_mean": 2.2387096881866455,
|
| 97 |
+
"moe_active": 2.8516128063201904,
|
| 98 |
+
"moe_aux": 3.1677749156951904
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"layer": 43,
|
| 102 |
+
"executed": 4,
|
| 103 |
+
"steps_mean": 2.59354829788208,
|
| 104 |
+
"moe_active": 3.3935482501983643,
|
| 105 |
+
"moe_aux": 3.4541893005371094
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"layer": 47,
|
| 109 |
+
"executed": 4,
|
| 110 |
+
"steps_mean": 1.6322580575942993,
|
| 111 |
+
"moe_active": 2.664515972137451,
|
| 112 |
+
"moe_aux": 3.1744186878204346
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"layer": 51,
|
| 116 |
+
"executed": 4,
|
| 117 |
+
"steps_mean": 1.793548345565796,
|
| 118 |
+
"moe_active": 2.8258063793182373,
|
| 119 |
+
"moe_aux": 3.864419460296631
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"layer": 55,
|
| 123 |
+
"executed": 4,
|
| 124 |
+
"steps_mean": 1.3096773624420166,
|
| 125 |
+
"moe_active": 2.2580645084381104,
|
| 126 |
+
"moe_aux": 3.897202491760254
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"layer": 59,
|
| 130 |
+
"executed": 4,
|
| 131 |
+
"steps_mean": 1.0903226137161255,
|
| 132 |
+
"moe_active": 3.245161294937134,
|
| 133 |
+
"moe_aux": 3.8669705390930176
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"layer": 63,
|
| 137 |
+
"executed": 3,
|
| 138 |
+
"steps_mean": 1.0322580337524414,
|
| 139 |
+
"moe_active": 2.8322579860687256,
|
| 140 |
+
"moe_aux": 5.9470534324646
|
| 141 |
+
}
|
| 142 |
+
],
|
| 143 |
+
"log_tail": [
|
| 144 |
+
{
|
| 145 |
+
"mode": "train",
|
| 146 |
+
"executed": [
|
| 147 |
+
2,
|
| 148 |
+
4,
|
| 149 |
+
4,
|
| 150 |
+
4,
|
| 151 |
+
4,
|
| 152 |
+
4,
|
| 153 |
+
4,
|
| 154 |
+
3
|
| 155 |
+
],
|
| 156 |
+
"steps_mean": 1.6166
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"mode": "train",
|
| 160 |
+
"executed": [
|
| 161 |
+
2,
|
| 162 |
+
4,
|
| 163 |
+
4,
|
| 164 |
+
4,
|
| 165 |
+
4,
|
| 166 |
+
4,
|
| 167 |
+
4,
|
| 168 |
+
3
|
| 169 |
+
],
|
| 170 |
+
"steps_mean": 1.6105
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"mode": "train",
|
| 174 |
+
"executed": [
|
| 175 |
+
2,
|
| 176 |
+
4,
|
| 177 |
+
4,
|
| 178 |
+
4,
|
| 179 |
+
4,
|
| 180 |
+
4,
|
| 181 |
+
4,
|
| 182 |
+
3
|
| 183 |
+
],
|
| 184 |
+
"steps_mean": 1.6435
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"mode": "train",
|
| 188 |
+
"executed": [
|
| 189 |
+
2,
|
| 190 |
+
4,
|
| 191 |
+
4,
|
| 192 |
+
4,
|
| 193 |
+
4,
|
| 194 |
+
4,
|
| 195 |
+
4,
|
| 196 |
+
3
|
| 197 |
+
],
|
| 198 |
+
"steps_mean": 1.564
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"mode": "train",
|
| 202 |
+
"executed": [
|
| 203 |
+
2,
|
| 204 |
+
4,
|
| 205 |
+
4,
|
| 206 |
+
4,
|
| 207 |
+
4,
|
| 208 |
+
4,
|
| 209 |
+
4,
|
| 210 |
+
3
|
| 211 |
+
],
|
| 212 |
+
"steps_mean": 1.6875
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"mode": "train",
|
| 216 |
+
"executed": [
|
| 217 |
+
3,
|
| 218 |
+
4,
|
| 219 |
+
4,
|
| 220 |
+
4,
|
| 221 |
+
4,
|
| 222 |
+
4,
|
| 223 |
+
4,
|
| 224 |
+
3
|
| 225 |
+
],
|
| 226 |
+
"steps_mean": 1.6344
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"mode": "train",
|
| 230 |
+
"executed": [
|
| 231 |
+
3,
|
| 232 |
+
4,
|
| 233 |
+
4,
|
| 234 |
+
4,
|
| 235 |
+
4,
|
| 236 |
+
4,
|
| 237 |
+
4,
|
| 238 |
+
3
|
| 239 |
+
],
|
| 240 |
+
"steps_mean": 1.6526
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"mode": "train",
|
| 244 |
+
"executed": [
|
| 245 |
+
2,
|
| 246 |
+
4,
|
| 247 |
+
4,
|
| 248 |
+
4,
|
| 249 |
+
4,
|
| 250 |
+
4,
|
| 251 |
+
4,
|
| 252 |
+
3
|
| 253 |
+
],
|
| 254 |
+
"steps_mean": 1.6492
|
| 255 |
+
}
|
| 256 |
+
]
|
| 257 |
+
}
|
| 258 |
+
}
|