f205be0ea7
1. cohere.py:347-348 - Fixed wrong variable names in QK normalization. Used `Q`/`K` but variables were named `Qn`/`Kn`. This caused NameError when `use_qk_norm=True` (e.g., c4ai-command-r-plus models). 2. cohere.py:482 - Fixed wrong object reference in inference loop. Used `self.mlp` but should be `decoder_layer.mlp` since we're iterating through decoder layers. Caused AttributeError during inference. 3. falcon_h1.py:459,461 - Fixed wrong attribute names in inference path. Used `post_attention_layernorm` and `mlp` but Falcon H1 uses `pre_ff_layernorm` and `feed_forward`. Caused AttributeError during generation. 4. qwen3_moe.py:210 - Fixed wrong module path with incorrect capitalization. Used `transformers.models.Qwen3Moe` but should be `transformers.models.qwen3_moe`. Caused AttributeError when patching rotary embeddings. 5. qwen3_moe.py:239 - Fixed wrong model_patcher class. Used `FastQwen3Model` but should be `FastQwen3MoeModel` for MoE models. Caused incorrect patching for Qwen3 MoE models. 6. hf_hub.py:21-22 - Fixed floor division and missing return for billion values. Used `//` instead of `/` for millions, and had no return for values >= 1B. Caused incorrect formatting and None return for large numbers. 7. save.py:550 - Fixed self-assignment that did nothing. `sharded_ram_usage = sharded_ram_usage` should be `= max_shard_size`. Caused integer shard sizes to be ignored. 8. rl.py:562-567 - Fixed orphan string not included in length_check. The elif branch for max_seq_length validation was a standalone string expression, not concatenated to length_check. Caused silent skip of the max_seq_length > model_max_seq_length warning. 9. granite.py:49-52 - Fixed wrong model name and version in error message. Said "Gemma2" and "4.42.3" but should be "Granite" and "4.45.0".
80 lines
2.1 KiB
Python
80 lines
2.1 KiB
Python
from huggingface_hub import HfApi, ModelInfo
|
|
|
|
_HFAPI: HfApi = None
|
|
|
|
POPULARITY_PROPERTIES = [
|
|
"downloads",
|
|
"downloadsAllTime",
|
|
"trendingScore",
|
|
"likes",
|
|
]
|
|
THOUSAND = 1000
|
|
MILLION = 1000000
|
|
BILLION = 1000000000
|
|
|
|
|
|
def formatted_int(value: int) -> str:
|
|
if value < THOUSAND:
|
|
return str(value)
|
|
elif value < MILLION:
|
|
return f"{float(value) / 1000:,.1f}K"
|
|
elif value < BILLION:
|
|
return f"{float(value) / 1000000:,.1f}M"
|
|
else:
|
|
return f"{float(value) / 1000000000:,.1f}B"
|
|
|
|
|
|
def get_model_info(
|
|
model_id: str, properties: list[str] = ["safetensors", "lastModified"]
|
|
) -> ModelInfo:
|
|
"""
|
|
Get the model info for a specific model.
|
|
|
|
properties: list[str] = See https://huggingface.co/docs/huggingface_hub/api-ref/hf_hub/hf_api/model_info
|
|
Default properties: ["safetensors", "lastModified"], only retrieves minimal information.
|
|
Set to None to retrieve the full model information.
|
|
"""
|
|
global _HFAPI
|
|
if _HFAPI is None:
|
|
_HFAPI = HfApi()
|
|
try:
|
|
model_info: ModelInfo = _HFAPI.model_info(model_id, expand = properties)
|
|
except Exception as e:
|
|
print(f"Error getting model info for {model_id}: {e}")
|
|
model_info = None
|
|
return model_info
|
|
|
|
|
|
def list_models(
|
|
properties: list[str] = None,
|
|
full: bool = False,
|
|
sort: str = "downloads",
|
|
author: str = "unsloth",
|
|
search: str = None,
|
|
limit: int = 10,
|
|
) -> list[ModelInfo]:
|
|
"""
|
|
Retrieve model information from the Hugging Face Hub.
|
|
|
|
properties: list[str] = See https://huggingface.co/docs/huggingface_hub/api-ref/hf_hub/hf_api/list_models
|
|
full: bool = Whether to retrieve the full model information, if True properties will be ignored.
|
|
sort: str = The sort order.
|
|
author: str = The author of the model.
|
|
search: str = The search query for filtering models.
|
|
|
|
"""
|
|
global _HFAPI
|
|
if _HFAPI is None:
|
|
_HFAPI = HfApi()
|
|
if full:
|
|
properties = None
|
|
|
|
models: list[ModelInfo] = _HFAPI.list_models(
|
|
author = author,
|
|
search = search,
|
|
sort = sort,
|
|
limit = limit,
|
|
expand = properties,
|
|
full = full,
|
|
)
|
|
return models
|