~Uh6ddlmZmZddlZddlZddlmZddlZddlm Z ddl m Z ddl m Z dZhdZd Zej"eZd ed eefd Zd eded efdZd eeeffdZd eefdZde d dfdZddZy))IterableIteratorN)Path)get_cuda_bnb_library_path) CUDASpecs)print_dedented) CONDA_PREFIXLD_LIBRARY_PATH>_PWDHOMEMAILPATHTMUXSHELLOLDPWDSSH_TTYLESSOPEN LESSCLOSE SSH_AUTH_SOCK XDG_DATA_DIRSXDG_RUNTIME_DIRXDG_GREETER_DATA_DIRDBUS_SESSION_BUS_ADDRESSGOOGLE_VM_CONFIG_LOCK_FILE)z cudart64*.dllzlibcudart*.so*z nvcuda*.dllpaths_list_candidatereturnc#K|jtjD]}|stj|vr t |} |j st jd|O tD]-}|j|D]}|js|/y#t$rYDwxYw#ttf$rYwxYww)Nz?The directory listed in your path is found to be non-existent: ) splitospathsepseprexistsloggerwarningOSErrorCUDA_RUNTIME_LIB_PATTERNSglobis_filePermissionError)r dir_stringdir lib_patternpths o/var/www/catia.catastroantioquia-mas.com/valormas/lib/python3.12/site-packages/bitsandbytes/diagnostics/cuda.py find_cuda_libraries_in_path_listr0+s*00<   66 #  z"C zz|NN%dehdi#jk$ 9 " 88K0"C{{}! " "   )   sX9C B:(B+0C1.B: B:(C+ B74B:6B77B::C  C C  Cenv_varvaluecr|tvxs.tj|vxr|tvxrd|vxr d|vxrd|vS)NCONDA BASH_FUNC )CUDART_PATH_PREFERRED_ENVVARSr r"CUDART_PATH_IGNORED_ENVVARSr1r2s r/is_relevant_candidate_env_varr:AsY00 FFeO ":: "w& "7* "E! ctjjDcic]\}}t||s||c}}Scc}}w)N)r environitemsr:r9s r/,get_potentially_lib_path_containing_env_varsr?Ns7131A1A1C u~wGdelnsGtGUN uu us AAc#Kt}tD]0}||vs||}t|Ed{|j|2|j D]\}}t|Ed{y7C7 w)a Searches for a cuda installations, in the following order of priority: 1. active conda env 2. LD_LIBRARY_PATH 3. any other env vars, while ignoring those that - are known to be unrelated - don't contain the path separator `/` If multiple libraries are found in part 3, we optimistically try one, while giving a warning message. N)r?r7r0popr>)candidate_env_varsenvvar directoryr1r2s r/find_cudart_librariesrERsFG/+ ' '*62I7 B B B  " "6 * + -224;3E:::; C ;s&A6A6A2;A6*A4+A64A6 cuda_specsctd|jd|jdt|}|j st d|d|j s t dyy)Nz%PyTorch settings found: CUDA_VERSION=z, Highest Compute Capability: .z Library not found: z8. Maybe you need to compile it from source? a WARNING: Compute capability < 7.5 detected! Only slow 8-bit matmul is supported for your GPU! If you run into issues with 8-bit matmul, you can try 4-bit quantization: https://huggingface.co/blog/4bit-transformers-bitsandbytes )printcuda_version_stringhighest_compute_capabilityrr#rhas_imma)rF binary_paths r/print_cuda_diagnosticsrNjs / 0N0N/OP''1'L'L&MQ P ,J7K     +}-         r;ctt}|s tdyt|dkDr=t dt j jd|D]}td|yy)NzLCUDA SETUP: WARNING! CUDA runtime files not found in any environmental path.z Found duplicate CUDA runtime files (see below). We select the PyTorch default CUDA runtime, which is a], but this might mismatch with the CUDA version that is needed for bitsandbytes. To override this behavior set the `BNB_CUDA_VERSION=` environmental variable. For example, if you want to use the CUDA version 122, BNB_CUDA_VERSION=122 python ... OR set the environmental variable in your .bashrc: export BNB_CUDA_VERSION=122 In the case of a manual override, make sure you set LD_LIBRARY_PATH, e.g. export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-11.2, z* Found CUDA runtime at: )listrErIlenrtorchversioncuda) cudart_pathsr.s r/print_cuda_runtime_diagnosticsrWs|-/0L  \] \ Q BCH--BTBTAU V   $  5C -cU3 4 5' r;)rN)collections.abcrrloggingr pathlibrrSbitsandbytes.cextensionrbitsandbytes.cuda_specsrbitsandbytes.diagnostics.utilsrr7r8r' getLogger__name__r$strr0boolr:dictr?rErNrWr;r/rds.  =-9 C(   8 $38D>, 3 s t vd38nv;x~;0 y T 25r;