github/tinygrad
1
1
Fork 0
mirror of https://github.com/tinygrad/tinygrad.git synced 2026-04-29 03:00:14 -04:00
Code Issues Packages Projects Releases Wiki Activity

load_gguf -> gguf_load in doc and test (#7199)

Browse Source
This commit is contained in:
chenyu
2024-10-21 14:03:33 -04:00
committed by GitHub
parent f93bd9e2b9
commit f37e6b453b
2 changed files with 7 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
test/unit/test_disk_tensor.py
View File

@@ -486,10 +486,10 @@ class TestGGUF(unittest.TestCase):
self.ctx = ctypes.cast(ggml.ggml_init(params), ctypes.POINTER(ctypes.c_void_p))
def tearDown(self) -> None: ggml.ggml_free(self.ctx)
def test_load_tinyllama_q8_0(self): self._test_load_gguf("https://huggingface.co/ggml-org/models/resolve/main/tinyllamas/stories15M-q8_0.gguf?download=true")
def test_load_tinyllama_q4_0(self): self._test_load_gguf("https://huggingface.co/ggml-org/models/resolve/main/tinyllamas/stories15M-q4_0.gguf?download=true")
def test_load_gpt2_q4_1(self): self._test_load_gguf("https://huggingface.co/PrunaAI/gpt2-GGUF-smashed/resolve/main/gpt2.Q4_1.gguf?download=true")
def test_load_sample_q6_k(self): self._test_load_gguf("https://huggingface.co/Isotr0py/test-gguf-sample/resolve/main/Quant_Q6_K_1024.gguf?download=true")
def test_load_tinyllama_q8_0(self): self._test_gguf_load("https://huggingface.co/ggml-org/models/resolve/main/tinyllamas/stories15M-q8_0.gguf?download=true")
def test_load_tinyllama_q4_0(self): self._test_gguf_load("https://huggingface.co/ggml-org/models/resolve/main/tinyllamas/stories15M-q4_0.gguf?download=true")
def test_load_gpt2_q4_1(self): self._test_gguf_load("https://huggingface.co/PrunaAI/gpt2-GGUF-smashed/resolve/main/gpt2.Q4_1.gguf?download=true")
def test_load_sample_q6_k(self): self._test_gguf_load("https://huggingface.co/Isotr0py/test-gguf-sample/resolve/main/Quant_Q6_K_1024.gguf?download=true")
def test_dequantization_q4_0(self): self._test_dequantization(ggml.GGML_TYPE_Q4_0)
def test_dequantization_q4_1(self): self._test_dequantization(ggml.GGML_TYPE_Q4_1)
@@ -514,7 +514,8 @@ class TestGGUF(unittest.TestCase):
dq_tensor = ggml_data_to_tensor(q_tensor, n_el, ttype).reshape(n_el)
np.testing.assert_equal(dq_tensor.numpy(), np.frombuffer(c_dq_data, dtype=np.float32))
def _test_load_gguf(self, url: str):
def _test_gguf_load(self, url: str):
fp = fetch(url)
model_size = os.stat(fp).st_size
gguf_tensor = Tensor.empty(model_size, dtype=dtypes.uint8, device=f"disk:{fp}").to(Device.DEFAULT)