{"id":24568,"library":"seahorse","title":"Seahorse","description":"Seahorse is a Pytorch-based library for constructing and training large-scale transformer language models, providing tools for distributed training, model parallelism, and efficient memory usage. Currently at version 1.2.3, with active development.","status":"active","version":"1.2.3","language":"python","source_language":"en","source_url":"https://github.com/corail-research/seahorse","tags":["deep-learning","transformers","pytorch","distributed-training"],"install":[{"cmd":"pip install seahorse","lang":"bash","label":"Latest stable release"}],"dependencies":[{"reason":"core dependency for tensor operations and neural networks","package":"torch","optional":false},{"reason":"used for model architectures and tokenizers","package":"transformers","optional":false}],"imports":[{"note":"Direct import of class without module path fails.","wrong":"import SeaModel","symbol":"SeaModel","correct":"from seahorse import SeaModel"}],"quickstart":{"code":"import torch\nfrom seahorse import SeaModel\nmodel = SeaModel.from_pretrained('seahorse-base')\ninputs = torch.randint(0, 1000, (1, 512))\noutputs = model(inputs)\nprint(outputs.shape)","lang":"python","description":"Load a pre-trained Seahorse model and run a forward pass."},"warnings":[{"fix":"Replace model_parallel=True with device_map='auto'.","message":"Breaking change: The 'model_parallel' argument was removed in v1.2.0. Use 'device_map' instead.","severity":"gotcha","affected_versions":">=1.2.0"},{"fix":"Replace load_checkpoint('path') with SeaModel.from_pretrained('path').","message":"The 'load_checkpoint' function is deprecated in v1.2.0; use 'SeaModel.from_pretrained' instead.","severity":"deprecated","affected_versions":">=1.2.0"},{"fix":"Verify Python version: python --version >= 3.11.","message":"Seahorse requires Python >=3.11; installing on older Python fails silently or with cryptic errors.","severity":"gotcha","affected_versions":"all"}],"env_vars":null,"last_verified":"2026-05-01T00:00:00.000Z","next_check":"2026-07-30T00:00:00.000Z","problems":[{"fix":"Use: from seahorse import SeaModel","cause":"Incorrect import path; seahorse exposes SeaModel at top level.","error":"ImportError: cannot import name 'SeaModel' from 'seahorse'"},{"fix":"Use device_map='auto' instead of model_parallel=True.","cause":"Removed in v1.2.0; replaced by device_map.","error":"ValueError: The argument `model_parallel` is not supported in this version."},{"fix":"Enable CPU offloading with device_map='sequential' or reduce batch size.","cause":"Model too large for GPU; need memory optimization or CPU offloading.","error":"RuntimeError: CUDA out of memory."},{"fix":"Use SeaModel.from_pretrained(path) instead.","cause":"load_checkpoint was removed from SeaModel instances; use class method.","error":"AttributeError: 'SeaModel' object has no attribute 'load_checkpoint'"},{"fix":"pip install torch","cause":"PyTorch not installed; required dependency.","error":"ModuleNotFoundError: No module named 'torch'"}],"ecosystem":"pypi","meta_description":null,"install_score":null,"install_tag":null,"quickstart_score":null,"quickstart_tag":null}