tocmo0nlord/axolotl
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Built site for gh-pages

Browse Source
This commit is contained in:
Quarto GHA Workflow Runner
2025-05-28 14:06:06 +00:00
parent 841f9f9b6c
commit 1211da1da2
4 changed files with 195 additions and 195 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
search.json
View File

@@ -2626,14 +2626,14 @@
"href": "docs/api/loaders.model.html",
"title": "loaders.model",
"section": "",
"text": "loaders.model\nModel loader class implementation for loading, configuring, and patching various\nmodels.\n\n\n\n\n\nName\nDescription\n\n\n\n\nModelLoader\nManages model configuration, initialization and application of patches during\n\n\n\n\n\nloaders.model.ModelLoader(\n self,\n cfg,\n tokenizer,\n *,\n inference=False,\n reference_model=False,\n **kwargs,\n)\nManages model configuration, initialization and application of patches during\nmodel loading.\nThis class orchestrates the entire process of loading a model from configuration to\nfinal preparation. It handles device mapping, quantization, attention mechanisms,\nadapter integration, and various optimizations.\n\n\n\nLoading and validating model configuration\nApplying monkey patches for optimizations / fixes\nSetting up device mapping (including multi-GPU configurations)\nConfiguring quantization\nSetting attention mechanisms (Flash Attention, SDPA, etc.)\nLoading and initializing the model\nApplying adapters (LoRA, QLoRA, etc.)\n\n\n\n\n\n\n\n\n\n\n\n\nName\nType\nDescription\n\n\n\n\nmodel\nPreTrainedModel | PeftModel | PeftMixedModel\nThe loaded model instance (available after load() is called).\n\n\nmodel_kwargs\ndict[str, Any]\nDictionary of keyword arguments passed to model initialization.\n\n\nbase_model\n\nName or path of the base model to load.\n\n\nmodel_type\n\nType of model to load (e.g., AutoModelForCausalLM).\n\n\nmodel_config\n\nConfiguration object for the model.\n\n\nauto_model_loader\n\nclass used for loading the model (default: AutoModelForCausalLM).\n\n\n\n\n\n\n\n\n\nName\nDescription\n\n\n\n\nload\nLoad and prepare the model with all configurations and patches.\n\n\n\n\n\nloaders.model.ModelLoader.load()\nLoad and prepare the model with all configurations and patches.\n\n\n\n\n\n\n\n\n\n\nName\nType\nDescription\n\n\n\n\n\ntuple[PreTrainedModel, PeftConfig | None]\nA tuple with the loaded model and its LoRA configuration (if applicable)."
"text": "loaders.model\nModel loader class implementation for loading, configuring, and patching various\nmodels.\n\n\n\n\n\nName\nDescription\n\n\n\n\nModelLoader\nManages model configuration, initialization and application of patches during\n\n\n\n\n\nloaders.model.ModelLoader(\n self,\n cfg,\n tokenizer,\n *,\n inference=False,\n reference_model=False,\n **kwargs,\n)\nManages model configuration, initialization and application of patches during\nmodel loading.\nThis class orchestrates the entire process of loading a model from configuration to\nfinal preparation. It handles device mapping, quantization, attention mechanisms,\nadapter integration, and various optimizations.\n\n\n\nLoading and validating model configuration\nApplying monkey patches for optimizations / fixes\nSetting up device mapping (including multi-GPU configurations)\nConfiguring quantization\nSetting attention mechanisms (Flash Attention, SDPA, etc.)\nLoading and initializing the model\nApplying adapters (LoRA, QLoRA, etc.)\n\n\n\n\n\n\n\n\n\n\n\n\nName\nType\nDescription\n\n\n\n\nmodel\nPreTrainedModel | PeftModel | PeftMixedModel\nThe loaded model instance (available after load() is called).\n\n\nmodel_kwargs\ndict[str, Any]\nDictionary of keyword arguments passed to model initialization.\n\n\nbase_model\n\nName or path of the base model to load.\n\n\nmodel_type\n\nType of model to load (e.g., AutoModelForCausalLM).\n\n\nmodel_config\n\nConfiguration object for the model.\n\n\nauto_model_loader\n\nclass used for loading the model (default: AutoModelForCausalLM).\n\n\n\n\n\n\n\n\n\nName\nDescription\n\n\n\n\nload\nLoad and prepare the model with all configurations and patches.\n\n\n\n\n\nloaders.model.ModelLoader.load()\nLoad and prepare the model with all configurations and patches.\n\n\n\n\n\n\n\n\n\n\nName\nType\nDescription\n\n\n\n\n\ntuple[PreTrainedModel | PeftModelForCausalLM, PeftConfig | None]\nA tuple with the loaded model and its LoRA configuration (if applicable)."
},
{
"objectID": "docs/api/loaders.model.html#classes",
"href": "docs/api/loaders.model.html#classes",
"title": "loaders.model",
"section": "",
"text": "Name\nDescription\n\n\n\n\nModelLoader\nManages model configuration, initialization and application of patches during\n\n\n\n\n\nloaders.model.ModelLoader(\n self,\n cfg,\n tokenizer,\n *,\n inference=False,\n reference_model=False,\n **kwargs,\n)\nManages model configuration, initialization and application of patches during\nmodel loading.\nThis class orchestrates the entire process of loading a model from configuration to\nfinal preparation. It handles device mapping, quantization, attention mechanisms,\nadapter integration, and various optimizations.\n\n\n\nLoading and validating model configuration\nApplying monkey patches for optimizations / fixes\nSetting up device mapping (including multi-GPU configurations)\nConfiguring quantization\nSetting attention mechanisms (Flash Attention, SDPA, etc.)\nLoading and initializing the model\nApplying adapters (LoRA, QLoRA, etc.)\n\n\n\n\n\n\n\n\n\n\n\n\nName\nType\nDescription\n\n\n\n\nmodel\nPreTrainedModel | PeftModel | PeftMixedModel\nThe loaded model instance (available after load() is called).\n\n\nmodel_kwargs\ndict[str, Any]\nDictionary of keyword arguments passed to model initialization.\n\n\nbase_model\n\nName or path of the base model to load.\n\n\nmodel_type\n\nType of model to load (e.g., AutoModelForCausalLM).\n\n\nmodel_config\n\nConfiguration object for the model.\n\n\nauto_model_loader\n\nclass used for loading the model (default: AutoModelForCausalLM).\n\n\n\n\n\n\n\n\n\nName\nDescription\n\n\n\n\nload\nLoad and prepare the model with all configurations and patches.\n\n\n\n\n\nloaders.model.ModelLoader.load()\nLoad and prepare the model with all configurations and patches.\n\n\n\n\n\n\n\n\n\n\nName\nType\nDescription\n\n\n\n\n\ntuple[PreTrainedModel, PeftConfig | None]\nA tuple with the loaded model and its LoRA configuration (if applicable)."
"text": "Name\nDescription\n\n\n\n\nModelLoader\nManages model configuration, initialization and application of patches during\n\n\n\n\n\nloaders.model.ModelLoader(\n self,\n cfg,\n tokenizer,\n *,\n inference=False,\n reference_model=False,\n **kwargs,\n)\nManages model configuration, initialization and application of patches during\nmodel loading.\nThis class orchestrates the entire process of loading a model from configuration to\nfinal preparation. It handles device mapping, quantization, attention mechanisms,\nadapter integration, and various optimizations.\n\n\n\nLoading and validating model configuration\nApplying monkey patches for optimizations / fixes\nSetting up device mapping (including multi-GPU configurations)\nConfiguring quantization\nSetting attention mechanisms (Flash Attention, SDPA, etc.)\nLoading and initializing the model\nApplying adapters (LoRA, QLoRA, etc.)\n\n\n\n\n\n\n\n\n\n\n\n\nName\nType\nDescription\n\n\n\n\nmodel\nPreTrainedModel | PeftModel | PeftMixedModel\nThe loaded model instance (available after load() is called).\n\n\nmodel_kwargs\ndict[str, Any]\nDictionary of keyword arguments passed to model initialization.\n\n\nbase_model\n\nName or path of the base model to load.\n\n\nmodel_type\n\nType of model to load (e.g., AutoModelForCausalLM).\n\n\nmodel_config\n\nConfiguration object for the model.\n\n\nauto_model_loader\n\nclass used for loading the model (default: AutoModelForCausalLM).\n\n\n\n\n\n\n\n\n\nName\nDescription\n\n\n\n\nload\nLoad and prepare the model with all configurations and patches.\n\n\n\n\n\nloaders.model.ModelLoader.load()\nLoad and prepare the model with all configurations and patches.\n\n\n\n\n\n\n\n\n\n\nName\nType\nDescription\n\n\n\n\n\ntuple[PreTrainedModel | PeftModelForCausalLM, PeftConfig | None]\nA tuple with the loaded model and its LoRA configuration (if applicable)."
},
{
"objectID": "docs/api/utils.bench.html",