Text Generation
Transformers
Safetensors
olmo3
custom-code
siamese-norm
depth-attention
sliding-window-attention
Instructions to use ArchSpace-Collection/SiameseNorm-DepthAttention with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ArchSpace-Collection/SiameseNorm-DepthAttention with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="ArchSpace-Collection/SiameseNorm-DepthAttention")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("ArchSpace-Collection/SiameseNorm-DepthAttention", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use ArchSpace-Collection/SiameseNorm-DepthAttention with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "ArchSpace-Collection/SiameseNorm-DepthAttention" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "ArchSpace-Collection/SiameseNorm-DepthAttention", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/ArchSpace-Collection/SiameseNorm-DepthAttention
- SGLang
How to use ArchSpace-Collection/SiameseNorm-DepthAttention 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 "ArchSpace-Collection/SiameseNorm-DepthAttention" \ --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": "ArchSpace-Collection/SiameseNorm-DepthAttention", "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 "ArchSpace-Collection/SiameseNorm-DepthAttention" \ --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": "ArchSpace-Collection/SiameseNorm-DepthAttention", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use ArchSpace-Collection/SiameseNorm-DepthAttention with Docker Model Runner:
docker model run hf.co/ArchSpace-Collection/SiameseNorm-DepthAttention
| { | |
| "architecture": "Olmo3SiameseDepthForCausalLM", | |
| "parameter_count": 1484716032, | |
| "recommended_attention_backend": "sdpa", | |
| "repo_id": "ArchSpace-Collection/SiameseNorm-DepthAttention", | |
| "schema_version": 2, | |
| "stages": { | |
| "stage1": { | |
| "checkpoint_iteration": 89407, | |
| "context_length": 8192, | |
| "subfolder": "olmo3/1b/stage1", | |
| "weight_bytes": 2969458432, | |
| "weight_sha256": "dea5562210667dc24ecec56159f05b9f8f2245aa83d64bf31e95dd02fff39695" | |
| }, | |
| "stage2": { | |
| "checkpoint_iteration": 47684, | |
| "context_length": 8192, | |
| "subfolder": "olmo3/1b/stage2", | |
| "weight_bytes": 2969458432, | |
| "weight_sha256": "226d1174f58fccc681a916163eabdfb7c5c9c15990fa4f7e4d3b88d721f9f849" | |
| }, | |
| "stage3": { | |
| "checkpoint_iteration": 11921, | |
| "context_length": 65536, | |
| "subfolder": "olmo3/1b/stage3", | |
| "weight_bytes": 2969458432, | |
| "weight_sha256": "0d380f9b5ff781d7c10089db0559215ad2da91443100d33e8e6fd677ff219622" | |
| }, | |
| "stage4_instruct": { | |
| "checkpoint_iteration": 3252, | |
| "context_length": 65536, | |
| "subfolder": "olmo3/1b/stage4/instruct", | |
| "weight_bytes": 2969458432, | |
| "weight_sha256": "f0b8b78613e1b49a5f6164acefec39e06f7f938176ff23663f703802799e2497" | |
| }, | |
| "stage4_think": { | |
| "checkpoint_iteration": 43224, | |
| "context_length": 65536, | |
| "subfolder": "olmo3/1b/stage4/think", | |
| "weight_bytes": 2969458432, | |
| "weight_sha256": "29c4567bcbcaf74acfca88df34c23fd23604f1ad66b933ff92bf72991c01ed58" | |
| } | |
| }, | |
| "state_tensor_count": 229, | |
| "validation": { | |
| "eager_backend": "loadable but not the BF16 release-validation backend; near-tied logits can expose cache-partition rounding differences", | |
| "hard_gate": "exact greedy token IDs and exact logical cache lengths", | |
| "npu_exact_logical_cache_growth": "pass", | |
| "npu_natural_prompt_32_token_three_path": "pass", | |
| "npu_readme_sdpa_all_stages": "pass", | |
| "npu_report_count": 16, | |
| "npu_sdpa_explicit_all_stages": "pass", | |
| "npu_stage3_65534_plus_2": "pass", | |
| "npu_stage4_chat_32_token_cache_parity": "pass", | |
| "npu_swa_and_yarn_boundaries": [ | |
| 4095, | |
| 4096, | |
| 4097, | |
| 8191, | |
| 8192, | |
| 8193, | |
| 65534 | |
| ], | |
| "numerical_metrics": "diagnostic only", | |
| "source_export_checksums_and_assembly": "pass" | |
| }, | |
| "vocab_size": 100278 | |
| } | |