fusion-bench 0.2.23__py3-none-any.whl → 0.2.24__py3-none-any.whl

fusion_bench/utils/devices.py

@@ -39,7 +39,12 @@ def clear_cuda_cache():
         log.warning("CUDA is not available. No cache to clear.")
 
 
-def to_device(obj: T, device: Optional[torch.device], **kwargs: Any) -> T:
+def to_device(
+    obj: T,
+    device: Optional[torch.device],
+    copy_on_move: bool = False,
+    **kwargs: Any,
+) -> T:
     """
     Move a given object to the specified device.
 
@@ -49,12 +54,20 @@ def to_device(obj: T, device: Optional[torch.device], **kwargs: Any) -> T:
     Args:
         obj: The object to be moved to the device. This can be a torch.Tensor, torch.nn.Module, list, tuple, or dict.
         device (torch.device): The target device to move the object to. This can be `None`.
-        **kwargs: Additional keyword arguments to be passed to the `to` method of torch.Tensor or torch.nn.Module. For example, `non_blocking=True`, `dtype=torch.float16`.
+        copy_on_move (bool, optional): Whether to force a copy operation when moving tensors to a different device.
+            If True, tensors will be copied when moved to a different device (copy=True is passed to tensor.to()).
+            If False (default), tensors are moved without forcing a copy operation, allowing PyTorch to optimize
+            the operation. This parameter only affects torch.Tensor objects; modules and other types are unaffected.
+            Defaults to False.
+        **kwargs: Additional keyword arguments to be passed to the `to` method of torch.Tensor or torch.nn.Module.
+            For example, `non_blocking=True`, `dtype=torch.float16`. Note that if `copy_on_move=True`, the `copy`
+            keyword argument will be automatically set and should not be provided manually.
 
     Returns:
         The object moved to the specified device. The type of the returned object matches the type of the input object.
 
     Examples:
+        ```python
         >>> tensor = torch.tensor([1, 2, 3])
         >>> to_device(tensor, torch.device('cuda'))
         tensor([1, 2, 3], device='cuda:0')
@@ -66,17 +79,26 @@ def to_device(obj: T, device: Optional[torch.device], **kwargs: Any) -> T:
         >>> data = [torch.tensor([1, 2]), torch.tensor([3, 4])]
         >>> to_device(data, torch.device('cuda'))
         [tensor([1, 2], device='cuda:0'), tensor([3, 4], device='cuda:0')]
+
+        >>> # Force copy when moving to different device
+        >>> tensor = torch.tensor([1, 2, 3], device='cpu')
+        >>> copied_tensor = to_device(tensor, torch.device('cuda'), copy_on_move=True)
+        >>> # tensor and copied_tensor will have different memory locations
+        ```
     """
-    if isinstance(obj, (torch.Tensor, torch.nn.Module)):
+    if isinstance(obj, torch.Tensor):
+        if copy_on_move:
+            if obj.device != torch.device(device):
+                kwargs["copy"] = True
+        return obj.to(device, **kwargs)
+    elif isinstance(obj, torch.nn.Module):
         return obj.to(device, **kwargs)
     elif isinstance(obj, list):
-        return [to_device(o, device) for o in obj]
+        return [to_device(o, device, **kwargs) for o in obj]
     elif isinstance(obj, tuple):
-        return tuple(to_device(o, device) for o in obj)
+        return tuple(to_device(o, device, **kwargs) for o in obj)
     elif isinstance(obj, dict):
-        for key in obj:
-            obj[key] = to_device(obj[key], device)
-        return obj
+        return {key: to_device(value, device, **kwargs) for key, value in obj.items()}
     else:
         # the default behavior is to return the object as is
         return obj

fusion_bench/utils/lazy_state_dict.py

@@ -76,6 +76,9 @@ class LazyStateDict(Mapping[str, torch.Tensor], Generic[TorchModelType]):
     _index: Optional[Dict[str, str]]
     """Mapping of parameter names to checkpoint files."""
 
+    meta_module: TorchModelType = None
+    meta_module_class: Optional[Type[TorchModelType]] = None
+
     def __init__(
         self,
         checkpoint: str,

fusion_bench/utils/rich_utils.py

@@ -188,17 +188,21 @@ if __name__ == "__main__":
     display_available_styles()
 
 
-def setup_colorlogging(force=False, **config_kwargs):
+def setup_colorlogging(
+    force=False,
+    level=logging.INFO,
+    **kwargs,
+):
     """
     Sets up color logging for the application.
     """
     FORMAT = "%(message)s"
 
     logging.basicConfig(
-        level=logging.INFO,
+        level=level,
         format=FORMAT,
         datefmt="[%X]",
         handlers=[RichHandler()],
         force=force,
-        **config_kwargs,
+        **kwargs,
     )

{fusion_bench-0.2.23.dist-info → fusion_bench-0.2.24.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fusion_bench
-Version: 0.2.23
+Version: 0.2.24
 Summary: A Comprehensive Benchmark of Deep Model Fusion
 Author-email: Anke Tang <tang.anke@foxmail.com>
 Project-URL: Repository, https://github.com/tanganke/fusion_bench
@@ -23,12 +23,19 @@ Requires-Dist: rich
 Requires-Dist: scipy
 Requires-Dist: h5py
 Requires-Dist: pytest
+Requires-Dist: joblib
+Requires-Dist: bidict
 Requires-Dist: transformers!=4.49
 Requires-Dist: pillow!=11.2.1
 Provides-Extra: lm-eval-harness
 Requires-Dist: lm-eval; extra == "lm-eval-harness"
 Requires-Dist: immutabledict; extra == "lm-eval-harness"
 Requires-Dist: langdetect; extra == "lm-eval-harness"
+Requires-Dist: rich-run; extra == "lm-eval-harness"
+Provides-Extra: docs
+Requires-Dist: mkdocs; extra == "docs"
+Requires-Dist: mkdocs-material; extra == "docs"
+Requires-Dist: mkdocstrings[python]; extra == "docs"
 Dynamic: license-file
 
 <div align='center'>
@@ -151,7 +158,7 @@ This will install the latest version of fusion-bench and the dependencies requir
 Documentation for using LM-Eval Harness within FusionBench framework can be found at [this online documentation](https://tanganke.github.io/fusion_bench/taskpool/lm_eval_harness) or in the [`docs/taskpool/lm_eval_harness.md`](docs/taskpool/lm_eval_harness.md) markdown file.
 
 > [!TIP]
-> Documentation for merging large language models using FusionBench can be found at [this online documentation](https://tanganke.github.io/fusion_bench/modelpool/causal_lm) or in the [`docs/modelpool/causal_lm.md`](docs/modelpool/causal_lm.md) markdown file.
+> Documentation for merging large language models using FusionBench can be found at [this online documentation](https://tanganke.github.io/fusion_bench/modelpool/llm) or in the [`docs/modelpool/llm/index.md`](docs/modelpool/llm/index.md) markdown file.
 
 ## Introduction to Deep Model Fusion
 
@@ -179,7 +186,7 @@ The project is structured as follows:
   - `taskpool`: configuration files for the task pool.
   - `model`: configuration files for the models.
   - `dataset`: configuration files for the datasets.
-- `docs/`: documentation for the benchmark. We use [mkdocs](https://www.mkdocs.org/) to generate the documentation. Start the documentation server locally with `mkdocs serve`. The required packages can be installed with `pip install -r mkdocs-requirements.txt`.
+- `docs/`: documentation for the benchmark. We use [mkdocs](https://www.mkdocs.org/) to generate the documentation. Start the documentation server locally with `mkdocs serve`. The required packages can be installed with `pip install -e ".[docs]"`.
 - `examples/`: example scripts for running some of the experiments.
   > **naming convention**: `examples/{method_name}/` contains the files such as bash scripts and jupyter notebooks for the specific method.
 - `tests/`: unit tests for the benchmark.

{fusion_bench-0.2.23.dist-info → fusion_bench-0.2.24.dist-info}/RECORD

@@ -48,12 +48,12 @@ fusion_bench/dataset/llama/stanford_shp.py,sha256=6ueXKnFXIBBobacU1h5WxGLZrSOtBk
 fusion_bench/dataset/llama/ultrachat.py,sha256=Go7WvrDAYnm184fdazHGRYLbSY6Xd7jrESyQeUJtOww,1736
 fusion_bench/dataset/llama/wikitext.py,sha256=9ZHR-nMfXRumd3o-PIj3n7B83YlVeqpGkZ2zJs2B-9Y,2883
 fusion_bench/dataset/llama/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-fusion_bench/method/__init__.py,sha256=MDYyNjJufoOe_iwmlL2ftWoD-72ReVv00mege5MQ6fc,8685
+fusion_bench/method/__init__.py,sha256=-d5WMlvY3kHYSUeompoG71T6fSttXPDjPf6X4TxNkqY,8986
 fusion_bench/method/base_algorithm.py,sha256=OnKSNPQ_nIdIWxryyblW_sko7uoEBN4lGh-eLkJ4kh4,9004
 fusion_bench/method/dummy.py,sha256=hb1y6LR_geRZ5eRgGwt5zJUcHYorCeIbs5i76CvurUc,1031
-fusion_bench/method/ensemble.py,sha256=oGiTJUderoPP0Opd7nHwC6h3VBmGTQ5inuG3wb6F4-A,3097
+fusion_bench/method/ensemble.py,sha256=Bjzqxt-tUp5cawT1jIhqKswN5QH3bkYbmuI4LS4uTG0,3619
 fusion_bench/method/model_recombination.py,sha256=b2ku5wCrWd1QSZscIra4KlhLDxt04JjU30ItMNvpZ6g,5268
-fusion_bench/method/simple_average.py,sha256=fLd14_0218JKyXmwe5M6kgumfD60u2ZVnm3B7PBX-Uc,5508
+fusion_bench/method/simple_average.py,sha256=FuIwHCUNK5CoToBzVt-lo8SK7wjj8CdRpiNLRnAflH4,5519
 fusion_bench/method/ada_svd/__init__.py,sha256=4XzQbbvE9HI3NtEmEFvo8iC3ds_85vJXe7P7qJfL7kk,77
 fusion_bench/method/ada_svd/clip_vision.py,sha256=XvXgIdlShAREMsubRgphyycGrhWqSnuVBo6S9bNYSd0,12581
 fusion_bench/method/adamerging/__init__.py,sha256=nt0saBT_3bqghk-pINQ-XCWm9UWwSZllu4R1sDuAJAA,376
@@ -128,12 +128,13 @@ fusion_bench/method/isotropic_merging/iso.py,sha256=MwKqfk0oyxqtdOzeSx_9jFXX1a4R
 fusion_bench/method/isotropic_merging/iso_utils.py,sha256=7L8PYUIJROwHJQmhFY-tdEhkLAnzVKXr-ae55FQ1QSo,6928
 fusion_bench/method/knots/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 fusion_bench/method/knots/knots_utils.py,sha256=NWlzo4nhQypUcNknU832MP3QT42VsLx-6WQ9QXuSigw,795
-fusion_bench/method/linear/__init__.py,sha256=ChfkoOEAb-rUKwpowFPel-a1hRfS8gCrbnWD-jlRbe4,283
+fusion_bench/method/linear/__init__.py,sha256=0U7JqlX8JuMECKlvLNM16Lxc1lCBN2bVqH8FtNoD-Fw,417
 fusion_bench/method/linear/expo.py,sha256=N7XnBTC0Nz_4gRs1f9TL9g-j-Lku5TF0lAjGKhZHwOw,3990
 fusion_bench/method/linear/linear_interpolation.py,sha256=Y01HPMBb7TaCjEBsbC6gqQyHvY1SRpwPyPPLxvYrL0s,2223
 fusion_bench/method/linear/llama_expo.py,sha256=ccECjhAqcFmzOIDyZ7e_aPzTM2Kj8u2D8TJytyz18YM,8476
-fusion_bench/method/linear/simple_average_for_llama.py,sha256=5psacdQiqtUK_lwYZcXp9kgIU3MFGk6G1JatxeMUjE8,3339
-fusion_bench/method/linear/task_arithmetic_for_llama.py,sha256=4SZpiTD7OzhWUXtcdK3PYdXbBGyDqiZd7oZOQ0lraN0,1963
+fusion_bench/method/linear/simple_average_for_causallm.py,sha256=qc-JiPLu19442DcP0xCl4EDGzVnIbq3WGiAiWkNzv6E,3448
+fusion_bench/method/linear/task_arithmetic_for_causallm.py,sha256=7cewnrjX47omokAdhNvDIQV8zz06_ZNKPWM7CZx30R0,2247
+fusion_bench/method/linear/ties_merging_for_causallm.py,sha256=yi0RCC6eRwXMKUC_cBdFLvejia4nmjPh9Pd0MpaUrVg,2392
 fusion_bench/method/lm_finetune/__init__.py,sha256=IFGAqXujX3Fabzl_tC6zZyOyPFJfVziL0qFtj5MVxj0,149
 fusion_bench/method/lm_finetune/bradley_terry_rm.py,sha256=1nvjOMABuEISyYaTRrFiwHLWvSTgHT8pEzTYBTLBRUg,18779
 fusion_bench/method/lm_finetune/causal_lm_pretrain.py,sha256=4CL9KGFsUzrt-edMfTooo4G4apzTH_57rso3DGGvKL0,219
@@ -222,7 +223,7 @@ fusion_bench/method/tall_mask/__init__.py,sha256=XINPP8PqGQ01he9p2RyHaKGyrcYoJuY
 fusion_bench/method/tall_mask/task_arithmetic.py,sha256=c-5ehKV_t46ljvKTBDr-eA3-FbSD_UNXlza4cOqK5aI,4371
 fusion_bench/method/tall_mask/utils.py,sha256=Wlp8WcPwR_lCaBIZ9rgG6ewLfSzz3G7kPk9yj13pvls,8817
 fusion_bench/method/task_arithmetic/__init__.py,sha256=pSx_NV5Ra_6UXpyYWCi6ANQoAnEtymZt_X1dDN9wT4Y,96
-fusion_bench/method/task_arithmetic/task_arithmetic.py,sha256=k4p8ADupDR5nZGHZjNgNsO8I_8rzqVyAr6Tejh85V0A,5525
+fusion_bench/method/task_arithmetic/task_arithmetic.py,sha256=KsSBshf04MUwIjoc0HAAmY6cWMqjZwZOYXbUuU4EaL0,6320
 fusion_bench/method/task_singular_vector/TSVC.py,sha256=yn4SrZNvtA6PoGYJmbmtNeDyDbGnRCgfZ7ZCg914AZU,410
 fusion_bench/method/task_singular_vector/TSVM.py,sha256=Sdgoi8xT0Hl19pmGdIuUS3D1DsVqSVD-Hipp-Sj_HoA,13652
 fusion_bench/method/task_singular_vector/__init__.py,sha256=WMucyl9pu_Ev2kcdrfT4moqMMbzD7hHQVFME5Su5jMA,298
@@ -231,7 +232,7 @@ fusion_bench/method/task_singular_vector/utils/TSVM_utils.py,sha256=WGM8wCICdGsN
 fusion_bench/method/task_singular_vector/utils/__init__.py,sha256=Mep62TnXJscBEFZ6QDsI28cWmfygt8EPwjQdfUJzEZQ,315
 fusion_bench/method/task_singular_vector/utils/task_singular_interference.py,sha256=tXsFwx8eomzu00nSp95CjjWZX82zq32ff2Q6VM_29CM,1348
 fusion_bench/method/ties_merging/__init__.py,sha256=9u9teBbdILbupr9jbwk-qCXSzssCssC5FUV2BfpyZM4,67
-fusion_bench/method/ties_merging/ties_merging.py,sha256=eCpGa9F4VoT0zsl7XKK7WsKz45tu_DkFHeffyJospJc,5152
+fusion_bench/method/ties_merging/ties_merging.py,sha256=u2o7Wo2SJJsxxhBeAhsmY7k4bdZkUtwAwGePGI4Sggc,5916
 fusion_bench/method/ties_merging/ties_merging_utils.py,sha256=EZyltS9hUM8NmcvXjAqhBpj-ucMlMtR95082kPDsJPU,10296
 fusion_bench/method/trust_region/__init__.py,sha256=4ao0E-jTlmTQPArbFWD_dFn_4yve3urNIuSMT8JtRIM,91
 fusion_bench/method/trust_region/clip_task_arithmetic.py,sha256=SWP7sRMiXzkDZ3KdNht3zqjaTcAtB4wpnnd8KYbcKZI,7441
@@ -245,6 +246,8 @@ fusion_bench/method/we_moe/we_moe.py,sha256=_QtmD04oFh7aLhmPq8EYchYB7BIN9ZFWOeys
 fusion_bench/method/weighted_average/__init__.py,sha256=bLxIuuB72hH05J_Spz4MZbiLpYL39iwgVIQa_QeQpIk,118
 fusion_bench/method/weighted_average/llama.py,sha256=vvxXp8v98kvXfHi7fYupnIrOVoA3tp08lmV2jDri_BY,3731
 fusion_bench/method/weighted_average/weighted_average.py,sha256=E4byEA2VfXozu7S_gnYVvwI3qg8AFWaSeNRHGbs2Tno,3340
+fusion_bench/method/wudi/__init__.py,sha256=08qPzOlhjw-Ab8TwyY9MGOGx_TLrUTueJc1WgRIvuxU,44
+fusion_bench/method/wudi/wudi.py,sha256=HL3Y0MPjozp7NML_UNjIWWPbQDQxYH_WG_BuyripeBQ,3602
 fusion_bench/metrics/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 fusion_bench/metrics/continual_learning/__init__.py,sha256=f-mkv4SpXTq5kiQVHbe2g0IPf4yLFgu1Dw7g2DOK6T4,57
 fusion_bench/metrics/continual_learning/backward_transfer.py,sha256=LCMWFFmBgWv7UIAJqiTaSvVvanx4qjnXIGuCMYvzmtc,559
@@ -262,10 +265,10 @@ fusion_bench/mixins/__init__.py,sha256=yjRvcB9Mn-c0g8tXmoBf2Dn8gyc-Na6dyhc4r674a
 fusion_bench/mixins/clip_classification.py,sha256=8dqJuI3AVetFZKuzTp1SR2kGQ-vGvfbcmwfnzuUiwfI,10096
 fusion_bench/mixins/fabric_training.py,sha256=ZmycEhCaNCgVi5oM9m0q6msxgk3quowmFvDAcvskFrg,13017
 fusion_bench/mixins/hydra_config.py,sha256=rfT-XPUKV_U3nvuTVsKLmSmEiieoSIsbhxE5_-E0er0,5508
-fusion_bench/mixins/lightning_fabric.py,sha256=ns9H_dkSDD8jJ7GL4YcAypewUcy9mzbX3Xy0bBcyGVY,7403
+fusion_bench/mixins/lightning_fabric.py,sha256=5iamAL7YV6lEm_-8NuzFjfIy1vslwKthSpCSWLLhlCM,7506
 fusion_bench/mixins/openclip_classification.py,sha256=O45HzgLXNvlQr5RVpfIGsYdIQ0tY5g_68KB0MTqsZWU,290
 fusion_bench/mixins/rich_live.py,sha256=j7wNgrgwfdpl6nCXZGF_2DLtNq2aqCb_52Qhe9QSltc,495
-fusion_bench/mixins/serialization.py,sha256=A2zEe3RIUhj60S8ENvjdMORz9zJ0bRnrAD54x1XIvao,15117
+fusion_bench/mixins/serialization.py,sha256=z73Mmq952TIdPwwZ8cRdl3n0_uc9lqylFI9fxKesREs,13260
 fusion_bench/mixins/simple_profiler.py,sha256=czWMl6p9PoxbQ5A8Uifwleaq5QPGEn0qMc8MXu9dSZM,2200
 fusion_bench/mixins/optim/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 fusion_bench/mixins/optim/adamw_with_warmup.py,sha256=qTnRl8GVVIfaplOFBHnJFuZUbxPZRWRGHGNzm_EDhDE,1421
@@ -277,7 +280,7 @@ fusion_bench/modelpool/huggingface_gpt2_classification.py,sha256=j8nicVwtoLXY4RP
 fusion_bench/modelpool/lazy_state_dict_pool.py,sha256=HtEA85rqSCHfsIddI5sKDcZf5kSuHNwrb8fF1TUSTr0,652
 fusion_bench/modelpool/nyuv2_modelpool.py,sha256=btuXmYxwfjI6MnGakhoOf53Iyb9fxYH20CavGTrTcnA,1375
 fusion_bench/modelpool/causal_lm/__init__.py,sha256=F432-aDIgAbUITj4GNZS9dgUKKhaDMCbTeHB-9MecaQ,99
-fusion_bench/modelpool/causal_lm/causal_lm.py,sha256=dSmjAhL4AxD34ckCdE8Rnf1hN5opoPIuz-hducQeK38,18685
+fusion_bench/modelpool/causal_lm/causal_lm.py,sha256=gpUQqxZIuKoaQ-gvdPsLVxI7UifueR6k3YzbUV1i0lk,19902
 fusion_bench/modelpool/clip_vision/__init__.py,sha256=3b9gN2bWUsoA1EmpitnIMnIlX7nklxbkn4WJ0QJtS2c,43
 fusion_bench/modelpool/clip_vision/modelpool.py,sha256=e5t9olRMOj_SyGVy-gqn7RwC5FAqxNsJDongWIv2KFY,7108
 fusion_bench/modelpool/openclip_vision/__init__.py,sha256=QDmAitKqUwRygN9QncdS_kGWZdfTKL4uUifC8xh9c10,47
@@ -288,8 +291,8 @@ fusion_bench/modelpool/seq_classification_lm/__init__.py,sha256=_VB9nlR_gm6IEXNM
 fusion_bench/modelpool/seq_classification_lm/reward_model.py,sha256=NKf-eoei1GdU4ojKSpN5_kQwax4uUEStnlKyh8qOrNg,540
 fusion_bench/modelpool/seq_classification_lm/seq_classification_lm.py,sha256=t9wXHFwa7V2XC3ajxt4_bSsxMTDKW4nebvdxhG7VeLM,3435
 fusion_bench/models/__init__.py,sha256=LeLQw2Yphu4QKZxjws_7MCM50XvFP1rTrvJ_2SR5zIA,271
-fusion_bench/models/hf_clip.py,sha256=056UHeSjKKDYXg-o7CC2zsx4fC9R6IBkPGI8IFhWTNw,7291
-fusion_bench/models/hf_utils.py,sha256=ozS56t69BOGy_wvbjX6MKFUuGsfKqy6s_TsinldNetk,5435
+fusion_bench/models/hf_clip.py,sha256=lL4LxbdwC_rDWRozdEJmRlzKaNcQMpWwCSMDE0tfZRM,7525
+fusion_bench/models/hf_utils.py,sha256=bfB3QAUqsG-TyUeOWrZt8V7GeWDhp-fKg3P0J3D_TbQ,5497
 fusion_bench/models/parameter_dict.py,sha256=HCkTJCz23pYN1_Hhegx8gglOtrnzVKJPMeg9_rUhe18,3630
 fusion_bench/models/rankone_moe.py,sha256=aY8IDM-ct7qKYH8ukBUsa_VDkDgGNtCqyNtNKlDTUTc,12046
 fusion_bench/models/separate_io.py,sha256=5AJlCxkHdVVffITnIRlF3ZIaKLRWDhJESVQN1lX-ZhU,3835
@@ -318,7 +321,7 @@ fusion_bench/models/llama/model_utils/mod.py,sha256=xzNOgTRfOK9q8kml4Q2nmSOl23f3
 fusion_bench/models/llama/model_utils/visual.py,sha256=wpqWqEASyA7WhJLCfC26h0Cdn5CXnwC1qPJUlSXggo4,8310
 fusion_bench/models/masks/__init__.py,sha256=vXG6jrBkDbPsnrX6nMEYAW1rQuGEWDgdjID7cKzXvrs,69
 fusion_bench/models/masks/mask_model.py,sha256=YXNZ_CGp6VPshZH__Znh6Z07BqOK53G-Ltc1LVy1E3I,5502
-fusion_bench/models/model_card_templates/default.md,sha256=Abd8tUhdZU-B5jwc7N6Gm0zLGNkfx6fr7MAL03VtFDg,885
+fusion_bench/models/model_card_templates/default.md,sha256=DJXwDODCsqIOhkgP57-iCShxLYK_jnsDsJYH1GfbBY8,1028
 fusion_bench/models/modeling_deepseek_v2/__init__.py,sha256=trXrhtKb_gIxXVo7wSZ-il5sLJtDTiNZezRrEt3M8zM,505
 fusion_bench/models/modeling_deepseek_v2/configuration_deepseek.py,sha256=TblFOCfNwaXUnXnD-sxFhSn5Df-_yy2LMcrth-sBPFI,10301
 fusion_bench/models/modeling_deepseek_v2/modeling_deepseek.py,sha256=PtfkfPrfmQVoLiVhgqlp5toJAnCinPWfeZYeJJtWWBs,78676
@@ -362,7 +365,7 @@ fusion_bench/models/smile_moe/utils/svd_utils.py,sha256=A2u7lH5Bo2qhgwplHPAz56pd
 fusion_bench/models/surgery/__init__.py,sha256=tcUSi2m9GzGWfvRDQScIbdEbFBS_35gm9zGKN7VpE70,53
 fusion_bench/models/surgery/surgerymodelwrapper.py,sha256=F8jX88K5zVWC6HsfN-nGNkEiPwNrN11ydyQQ1EZHehM,5133
 fusion_bench/models/wrappers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-fusion_bench/models/wrappers/ensemble.py,sha256=MQ92yxI_D8AzzA8sbpZE-rp-vWxO0tTICFnF8Y1Gyss,6380
+fusion_bench/models/wrappers/ensemble.py,sha256=T-DAKrAm-ciZwV6Hbt8uASbjtoQpHTlvVyan3rhk_8k,11632
 fusion_bench/models/wrappers/layer_wise_fusion.py,sha256=A7LjG0inL5oeEVOkJwEUDM15v4dpQnsCq2y9zA78R3k,11198
 fusion_bench/models/wrappers/layer_wise_fusion_doge_ta.py,sha256=q5Hc4BtLpAawMbxsWJRL-8OR-x7994Jhr9IyN7vKZ9o,16930
 fusion_bench/models/wrappers/task_wise_fusion.py,sha256=ROLANdDq0bZ3sIROqIv3udPN8lzDdEwxD0Jonx-5ycw,17465
@@ -377,7 +380,7 @@ fusion_bench/programs/__init__.py,sha256=oGoRp2TMI6ELxyfkeTg2h27hZJEDz9x31Asmvwv
 fusion_bench/programs/base_program.py,sha256=Bl_bv8SawEUc-GBTtZFMoii0y-r-0hOXBAJkQFexWCU,3475
 fusion_bench/programs/fabric_fusion_program.py,sha256=jt0_tlg37a2jBl2YikaC0N71Gmr4J340wkKAekyT180,12453
 fusion_bench/scripts/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-fusion_bench/scripts/cli.py,sha256=VwcwqY--kGDEGI1RoTQ5X32FaKducdRUKf2CZRXcfCM,2739
+fusion_bench/scripts/cli.py,sha256=kEWLEkZEBqUr1_-XTePzNC5NM8lwWvgUBf0Lcuk_FI8,2739
 fusion_bench/scripts/imgui.py,sha256=r9Glbfbwu3JCsX9TKQFwcHarvwA_G7ff0jWBUPW1S1U,7613
 fusion_bench/scripts/nyuv2_mtl_train.py,sha256=W1C45R9NdF4O-UjCx1bUxRTdFE0-FlRpwJHZ5gY18rI,3602
 fusion_bench/scripts/webui.py,sha256=ryA-2leSnHcYA88tTAYzJGDhiljbi0vl1Fibejzndlw,14398
@@ -392,7 +395,7 @@ fusion_bench/taskpool/clip_vision/__init__.py,sha256=ItdyWYy2A5xQKzh1dXi9kbQTBig
 fusion_bench/taskpool/clip_vision/clip_rankone_moe_taskpool.py,sha256=t_lmo8W-ZgLLOiBnF5CWfaLbKwz3EXfO8gCavI34qQY,3733
 fusion_bench/taskpool/clip_vision/clip_smile_taskpool.py,sha256=UdI7npI53LjPV2B19tHymhbma6WYcZIvzhqaSyZKkSQ,4762
 fusion_bench/taskpool/clip_vision/clip_sparse_wemoe_taskpool.py,sha256=8lZIG6tWpctYzme0Q_n6QcGnn9MeDmP3UX8nEv4_a9Q,4232
-fusion_bench/taskpool/clip_vision/taskpool.py,sha256=3JPN_1B9ylG0-Q69UELdQgakrgxRRQbj9x6LvTlw_J0,16177
+fusion_bench/taskpool/clip_vision/taskpool.py,sha256=99F8w_e4-UnoeDkSjo0z_8Wstx6e635h0IqSdtfT7ms,16460
 fusion_bench/taskpool/clip_vision/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 fusion_bench/taskpool/clip_vision/utils/routing_analysis_utils.py,sha256=LY9wxWCm_4X7Ii0ZkMxhtbevz6OxS3Bkqz0puXhuRqM,2393
 fusion_bench/taskpool/llama/__init__.py,sha256=iB4ESMgnsl0m-z0YtRdPZiwGGv96-86R8pbSnkdet8Q,57
@@ -442,7 +445,7 @@ fusion_bench/utils/__init__.py,sha256=wNAfpP-u_-8HGbLaBoHT_wriU_cNvY4M_UXdBv2kXh
 fusion_bench/utils/auto.py,sha256=uACQLE62_kNyhl4BGduvcbyeTE61qXpIJx3Ccl8kh68,920
 fusion_bench/utils/cache_utils.py,sha256=-bTZijQgl4BuAx0VSJFD-bSDOXuq3o0NkrOaiLiyofU,4795
 fusion_bench/utils/data.py,sha256=aalB3kGbZUF-PZ_IaAhcXanRKhS-RNMT5mUrEBb4R3E,6722
-fusion_bench/utils/devices.py,sha256=i5g2FzFs-UWhekcwzxVUZBOw82pOP-RbjIISbfWnuoM,8357
+fusion_bench/utils/devices.py,sha256=6AkGcs3flt0FSo9yfEREuehoTrgcc65gkwpTWQy8XsI,9546
 fusion_bench/utils/dict.py,sha256=ZCK0CRRT_B1Z18WY_GOYcmth7k5x9Jn1k7XhAVWRu98,1379
 fusion_bench/utils/dtype.py,sha256=z6UlPGF9dzG4Ik8rXGf59PJk_RKzG6Trp8O6wcBS9PU,4360
 fusion_bench/utils/expr.py,sha256=zwHNrtIbOMnIChU-0ZI5qLbDva8zvHbizL-4F2TwM14,2386
@@ -452,14 +455,14 @@ fusion_bench/utils/hydra_utils.py,sha256=TklUDKDEZlg4keI-TEZiqh4gFjr9-61Rt1RMlqk
 fusion_bench/utils/instantiate_utils.py,sha256=OXkfhq_o3Sgy5n3Psf-HI-dIfbK9oD2GBdfcx3gT63Q,17526
 fusion_bench/utils/json.py,sha256=sVCqbm9mmyHybiui-O57KFt_ULrjLtN2wipSo6VDvqE,2533
 fusion_bench/utils/lazy_imports.py,sha256=dg4Uu8FaoEu0WGVTo5o_PbLZs3Ei_RG75Ta-Us1iPW4,3500
-fusion_bench/utils/lazy_state_dict.py,sha256=9dse7U3QZNvNxBINb02Q9DW2_-voUh2Ri2B6hk9wvNI,20227
+fusion_bench/utils/lazy_state_dict.py,sha256=srEKyctbuBW3yrVFSG7Tki_XkBwoc6eUmDXLxHXqX0o,20328
 fusion_bench/utils/misc.py,sha256=93q0m-HYWkPK91Co5lll_J0Dxs6YahW2lD_X8fUAyTk,2420
 fusion_bench/utils/modelscope.py,sha256=P8fV6Eff8oP0LVGIFGbLvuk8MBteysN438djZ6ZEfE4,10699
 fusion_bench/utils/packages.py,sha256=wKl-qtPjA61LrdgTTusuNyvs8jfUv4mA5IwPTFWyYtA,2139
 fusion_bench/utils/parameters.py,sha256=ufEDOYJwcQQxLfveK8hBAGwpu5J3LA_cTWiDgZ2zkJ0,11788
 fusion_bench/utils/path.py,sha256=qrfgar3b-6_2v032-2hTt97L6qdtG7zc3CFrGFyKSGE,2400
 fusion_bench/utils/pylogger.py,sha256=r2KXTvq-j8uHdjBBoVPOgkjv4c6pyhbX6xf1JbOsF4w,3335
-fusion_bench/utils/rich_utils.py,sha256=XNPUpa1grna_C0MLQs0nY25-Kfutpj9BOEzvjoH7nR0,5849
+fusion_bench/utils/rich_utils.py,sha256=24RF-OHK6h9ggZ95csw_vMU8YtxYNOxlzjcH7dpuESY,5863
 fusion_bench/utils/set.py,sha256=_43ZvGKJ_BK9sUslsSNhi7xEfuAQuyj3vViImnGpnCY,134
 fusion_bench/utils/state_dict_arithmetic.py,sha256=fczHDEpL2_UmxNIdvQtllXvBWBcmKpw-p6CIS_upjwI,11818
 fusion_bench/utils/tensorboard.py,sha256=9fkgNYR9LM38nPNkudcxL9TjLUseW-280M0k2nLff7o,1669
@@ -472,7 +475,7 @@ fusion_bench/utils/plot/token_notebook.py,sha256=bsntXf46Zz_RavTxNiB9c3-KvHw7LFw
 fusion_bench/utils/strenum/__init__.py,sha256=id9ORi1uXrDxhbmVxitJ1KDwLS4H3AAwFpaK5h1cQzw,8531
 fusion_bench/utils/strenum/_name_mangler.py,sha256=o11M5-bURW2RBvRTYXFQIPNeqLzburdoWLIqk8X3ydw,3397
 fusion_bench/utils/strenum/_version.py,sha256=6JQRo9LcvODbCOeVFYQb9HNJ_J9XiG_Zbn8ws2A3BV8,18466
-fusion_bench-0.2.23.dist-info/licenses/LICENSE,sha256=nhnOJlw4CPuPVE0qvkGmxfFgHmKi-6nzXvTu8t0NUdg,1066
+fusion_bench-0.2.24.dist-info/licenses/LICENSE,sha256=nhnOJlw4CPuPVE0qvkGmxfFgHmKi-6nzXvTu8t0NUdg,1066
 fusion_bench_config/README.md,sha256=Lc8YSBJ5oxf9KV5kKDivJ9LRyGuraGQPmBbgbdVA-j4,703
 fusion_bench_config/clip-vit-base-patch32_robustness_corrupted.yaml,sha256=7IxLQoLRz-sRWyV8Vqc5kQcmYE_9YQz2_77pmvAkum8,1207
 fusion_bench_config/fabric_model_fusion.yaml,sha256=U8BxsaOvsg9bsEZcIpBE-feo9n9G7Y1kQDHqPVxUYAg,2601
@@ -621,7 +624,7 @@ fusion_bench_config/method/dare/ties_merging.yaml,sha256=7gDW4XpezrsccsbJGqqKrbX
 fusion_bench_config/method/dawe/dawe_for_clip.yaml,sha256=99P5xpp1YGvIwXGxDcxRtJMLE2FhvEFmFBQjOMEcGoc,1023
 fusion_bench_config/method/doge_ta/doge_ta.yaml,sha256=CtZI3YPMJNDy225yhOJbSiMKlsc-X5nCFzmVh0dvr-w,78
 fusion_bench_config/method/ensemble/max_model_predictor.yaml,sha256=khdpCvKMNytx4nZSgtUJFXv44MVytXu0aqUVd9TixXo,57
-fusion_bench_config/method/ensemble/simple_ensemble.yaml,sha256=Ih9dqifpnvxW2QfJqp8Q8S8W1k7VZG9ulyPxkcuaWsw,54
+fusion_bench_config/method/ensemble/simple_ensemble.yaml,sha256=RKa3IgN3DfFZVmeXVIdTt0NdPVV0jFkpQz6SxLs3Kso,124
 fusion_bench_config/method/ensemble/weighted_ensemble.yaml,sha256=2KD3PjFglqL7fjqhjXtOWxZ1mvmYodiNVroXsFd7EGE,261
 fusion_bench_config/method/expert_sparsity/README.md,sha256=CLE0-XblXDWCUTHPaTNtBH-YquXn-uawwTJiYrgjMaA,239
 fusion_bench_config/method/expert_sparsity/mixtral.yaml,sha256=maFL3LM0zfnQ1eXoNXUslSjgZmpOdUJgl_a31dYUBbc,605
@@ -638,8 +641,9 @@ fusion_bench_config/method/linear/expo.yaml,sha256=St3NW6cKVRV3vCn8y0gxQ8k66VTdt
 fusion_bench_config/method/linear/linear_interpolation.yaml,sha256=chM6_HRKKcMleTeuKY3-YNI1qaMG2CfnsRwUxAlHsRw,66
 fusion_bench_config/method/linear/llama_expo.yaml,sha256=SvqamjT06BMObQ58sks5x7Wv6kGpp3-Nlw3ihbD_kSA,621
 fusion_bench_config/method/linear/llama_expo_with_dare.yaml,sha256=Pp8s2xmEg5XSvaGKtwTYx_PzcGvwRh2gPpZ6u9as4_E,383
-fusion_bench_config/method/linear/simple_average_for_llama.yaml,sha256=r2Zul2GaMEEQ7NEDf8yiAgEiMDPNibU4qsJ0toD2KjQ,319
-fusion_bench_config/method/linear/task_arithmetic_for_llama.yaml,sha256=N7cyHm6a2QwNsV9uaJp-eZmdbs9kmdRrkxtO58QQQgM,116
+fusion_bench_config/method/linear/simple_average_for_causallm.yaml,sha256=qqeIr61PJEcfZclZ5vV64GCzyt-8b1zB0FDZu8DsbXQ,322
+fusion_bench_config/method/linear/task_arithmetic_for_causallm.yaml,sha256=tJA0n0_XVvll4rZYVHQVqFCz8W3Bey6NjPKMIH3-P0U,142
+fusion_bench_config/method/linear/ties_merging_for_causallm.yaml,sha256=1oEIdxV0OqWjDQ9V_lmXEPUayp4KbKHE2SvpCLmiKOU,489
 fusion_bench_config/method/linear/weighted_average.yaml,sha256=uq2gHGCwVHHSa1H-hzcrSlumUTLJ50tfyiY1Mh1pFsk,186
 fusion_bench_config/method/linear/weighted_average_for_llama.yaml,sha256=se2aq6t5R1f-ZG6ubUyRr__DBe9BzXrgL81ua3DkQoM,498
 fusion_bench_config/method/lm_finetune/bradley_terry_rm.yaml,sha256=QHsRfJK9K4KajsX3LBHG8cDt7ZLJWxOBnJjpHRQSB_s,1348
@@ -686,6 +690,7 @@ fusion_bench_config/method/trust_region/clip_task_arithmetic.yaml,sha256=-Ipc05T
 fusion_bench_config/method/wemoe/flan_t5_weight_ensembling_moe.yaml,sha256=KIKUr_Q4e9pJSVlqUFatuLp5vg8kNEsn8tOE4R77sxA,653
 fusion_bench_config/method/wemoe/sparse_weight_ensembling_moe.yaml,sha256=mMVaFJWUZmIdhg0kVQY20i7cmgTMrOSgoSpmW7quRzc,993
 fusion_bench_config/method/wemoe/weight_ensembling_moe.yaml,sha256=OEv5yhyUCe5lXeT2PyXC49yrHXEM7i8SZDw6IQRDtAE,620
+fusion_bench_config/method/wudi/wudi.yaml,sha256=3mJ6-XKHwwHALS3d503ybGM7pc1PhEK91YwwMybuzMc,76
 fusion_bench_config/model/clip-vit/README.md,sha256=-s34C9X7pxy55xSc24kbf-4ctK7UC-Wpu_JWIe9O0Ko,1382
 fusion_bench_config/model/clip-vit/clip-vit-base-patch16.yaml,sha256=Fn7or7-5fVZNyp6fH1lkwk7mq7iVhpR3sMt6Sm7Yg6I,43
 fusion_bench_config/model/clip-vit/clip-vit-base-patch16_TALL14.yaml,sha256=8G2OCCDaSJkzDOMDsV08NE-Z5YWMjDsFVs1WY3OWNss,787
@@ -841,9 +846,11 @@ fusion_bench_config/modelpool/CLIPVisionModelPool/clip-vit-large-patch14_TALL14_
 fusion_bench_config/modelpool/CLIPVisionModelPool/clip-vit-large-patch14_TALL20.yaml,sha256=FuPWQbC9xEV5wZjuo835gOMNgbzmpK9RbjFjA_HOzqo,2476
 fusion_bench_config/modelpool/CLIPVisionModelPool/clip-vit-large-patch14_TALL20_model_only.yaml,sha256=9PCkbrNnQSKTsm4eoUvVgjGd3IY7wHBC4LWj4kOdY4Y,1406
 fusion_bench_config/modelpool/CLIPVisionModelPool/clip-vit-large-patch14_individual.yaml,sha256=bqnyzgwIvDtV3Fb-uLf9mdFv0NW1C392lxGsGUPLsKE,400
-fusion_bench_config/modelpool/CausalLMPool/Qwen2.5-1.5B_math_and_coder.yaml,sha256=D8HdBRGUYD-c-c38oSgzcP3fkNhBN-tVdqLnS_B-7zc,265
+fusion_bench_config/modelpool/CausalLMPool/Qwen2.5-1.5B_math_and_code.yaml,sha256=4DoMFlGabtwZXZMGWsWtkP2rlGOx_1eEPp_AyqyVln0,263
+fusion_bench_config/modelpool/CausalLMPool/Qwen2.5-1.5B_three_models.yaml,sha256=ofFFVYKHKtylxd90REMLhhP57Yqwe2AEbGuZ0mBCVz8,305
 fusion_bench_config/modelpool/CausalLMPool/Qwen2.5-7B-math_and_coder.yaml,sha256=Nxk72MurqSzEyPJzGoKFbk5T2TGWBwYpH2V9Jzqt648,229
 fusion_bench_config/modelpool/CausalLMPool/deepseek-v2-lite.yaml,sha256=8gr8ZtgegSHV0GHtJBiEgdYbRe8UHhO4_y8dayxZChk,506
+fusion_bench_config/modelpool/CausalLMPool/llama-7b_3-models_v1.yaml,sha256=mm7A3NilcANJBuCZMt3MMLKFm7CjBhMYWAa9TXjM_PQ,326
 fusion_bench_config/modelpool/CausalLMPool/llama_alpaca_cleaned.yaml,sha256=oDsZkuAoh1mWUC7jZNzw8794zgX2bV5Z0esXpvbTs-c,643
 fusion_bench_config/modelpool/CausalLMPool/llama_codealpaca.yaml,sha256=FuUsBrvk3_bQiciMRlNsO5vp6AKHQM_-g-8bmU8251w,641
 fusion_bench_config/modelpool/CausalLMPool/llama_for_causallm.yaml,sha256=H3UwSk4ChsGSrH49LuttxldFURW-4RVUtnIa0ClHKXo,802
@@ -927,8 +934,8 @@ fusion_bench_config/taskpool/LMEvalHarnessTaskPool/lm_eval.yaml,sha256=3q-KMuFaM
 fusion_bench_config/taskpool/OpenCLIPVisionModelTaskPool/ViT-B-16_TA8.yaml,sha256=GjpiiRownrBCpl-TNwWRW2PYePbF-Cl99jlLNPrK5T4,1017
 fusion_bench_config/taskpool/OpenCLIPVisionModelTaskPool/ViT-B-32_TA8.yaml,sha256=WwiYMQKehtJixDPnu5o3vcWe4yJksXTWRqOzm3uVWXQ,1017
 fusion_bench_config/taskpool/OpenCLIPVisionModelTaskPool/ViT-L-14_TA8.yaml,sha256=xGRt0J9joXTzWUew6DvoYprAWlPXhaVFw5AX4im5VQw,1017
-fusion_bench-0.2.23.dist-info/METADATA,sha256=HQZ3DxHk-Jtcj2AZT49tx5m7VdVkDEglivhkfQv258Q,22384
-fusion_bench-0.2.23.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-fusion_bench-0.2.23.dist-info/entry_points.txt,sha256=iUQ8MCJvda7HP4vYh2n1Teoapb4G9PBVYZkAfcc5SHU,116
-fusion_bench-0.2.23.dist-info/top_level.txt,sha256=BuO4TL6iHL_2yPBUX9-LlIrHRczA_BNMIFwweK0PQEI,13
-fusion_bench-0.2.23.dist-info/RECORD,,
+fusion_bench-0.2.24.dist-info/METADATA,sha256=DllRpMnvVgyeqjN_YlNeo7IlqukzOjuYO_cWopOo1tA,22621
+fusion_bench-0.2.24.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+fusion_bench-0.2.24.dist-info/entry_points.txt,sha256=iUQ8MCJvda7HP4vYh2n1Teoapb4G9PBVYZkAfcc5SHU,116
+fusion_bench-0.2.24.dist-info/top_level.txt,sha256=BuO4TL6iHL_2yPBUX9-LlIrHRczA_BNMIFwweK0PQEI,13
+fusion_bench-0.2.24.dist-info/RECORD,,

fusion_bench_config/method/ensemble/simple_ensemble.yaml

@@ -1 +1,2 @@
 _target_: fusion_bench.method.SimpleEnsembleAlgorithm
+device_map: null  # Set to null for single device, or specify mapping

fusion_bench_config/method/linear/{simple_average_for_llama.yaml → simple_average_for_causallm.yaml}

@@ -1,4 +1,4 @@
-_target_: fusion_bench.method.SimpleAverageForLlama
+_target_: fusion_bench.method.SimpleAverageForCausalLM
 # set `merge_backbone` to true if you has a base model and only want to merge the backbone of the experts
 # if `merge_backbone` is False, this is equivalent to `SimpleAverageAlgorithm`
 merge_backbone: false

fusion_bench_config/method/linear/task_arithmetic_for_causallm.yaml

@@ -0,0 +1,4 @@
+_target_: fusion_bench.method.TaskArithmeticForCausalLM
+scaling_factor: 0.3
+merge_backbone: false
+model_save_path: ${path.log_dir}/checkpoint

fusion_bench_config/method/linear/ties_merging_for_causallm.yaml

@@ -0,0 +1,13 @@
+_target_: fusion_bench.method.TiesMergingForCausalLM
+# Scaling factor $\lambda$ 
+scaling_factor: 0.3
+# Threshold for resetting values in the task vector
+threshold: 20
+# List of keys to remove from the state dict, default is empty
+remove_keys: []
+# Function to merge the models, default is sum. Options are 'sum', 'mean', and 'max'
+merge_func: sum
+# Whether to merge only the backbone layers
+merge_backbone: false
+# Path to save the merged model
+model_save_path: ${path.log_dir}/checkpoint

fusion_bench_config/method/wudi/wudi.yaml

@@ -0,0 +1,4 @@
+_target_: fusion_bench.method.WUDIMerging
+
+iter_num: 400
+exclude_keys: null

fusion_bench_config/modelpool/CausalLMPool/{Qwen2.5-1.5B_math_and_coder.yaml → Qwen2.5-1.5B_math_and_code.yaml}

@@ -1,7 +1,6 @@
 _target_: fusion_bench.modelpool.CausalLMPool
 _recursive_: false
-
-enable_lazy_loading: false
+enable_lazy_loading: true
 models:
   _pretrained_: Qwen/Qwen2.5-1.5B
   math: Qwen/Qwen2.5-Math-1.5B

fusion_bench_config/modelpool/CausalLMPool/Qwen2.5-1.5B_three_models.yaml

@@ -0,0 +1,11 @@
+_target_: fusion_bench.modelpool.CausalLMPool
+_recursive_: false
+enable_lazy_loading: true
+models:
+  _pretrained_: Qwen/Qwen2.5-1.5B
+  math: Qwen/Qwen2.5-Math-1.5B
+  code: Qwen/Qwen2.5-Coder-1.5B
+  instruction: Qwen/Qwen2.5-1.5B-Instruct
+model_kwargs:
+  torch_dtype: bfloat16
+tokenizer: Qwen/Qwen2.5-1.5B

fusion_bench_config/modelpool/CausalLMPool/llama-7b_3-models_v1.yaml

@@ -0,0 +1,11 @@
+_target_: fusion_bench.modelpool.CausalLMPool
+_recursive_: false
+enable_lazy_loading: true
+models:
+  _pretrained_: meta-llama/Llama-2-7b-hf
+  chat: meta-llama/Llama-2-7b-chat-hf
+  math: WizardLMTeam/WizardMath-7B-V1.0
+  code: codellama/CodeLlama-7b-hf
+model_kwargs:
+  torch_dtype: bfloat16
+tokenizer: meta-llama/Llama-2-7b-hf

fusion_bench_config/method/linear/task_arithmetic_for_llama.yaml

@@ -1,4 +0,0 @@
-_target_: fusion_bench.method.TaskArithmeticForLlama
-scaling_factor: 0.3
-merge_backbone: true
-model_save_path: null