mirrors/docling
1
0
Fork 0
mirror of https://github.com/DS4SD/docling.git synced 2025-12-13 07:08:19 +00:00
Code Issues Packages Projects Releases Wiki Activity

pixtral 12b runs via MLX and native transformers

Browse Source 
Signed-off-by: Peter Staar <taa@zurich.ibm.com>
This commit is contained in:
Peter Staar
2025-05-14 07:39:20 +02:00
parent 054e01d8b3
commit f159075b67
2 changed files with 14 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
docs/examples/minimal_vlm_pipeline.py
View File

@@ -49,6 +49,7 @@ pixtral_vlm_conversion_options = HuggingFaceVlmOptions(
vlm_conversion_options = pixtral_vlm_conversion_options
"""
"""
pixtral_vlm_conversion_options = HuggingFaceVlmOptions(
repo_id="mistral-community/pixtral-12b",
prompt="OCR this image and export it in MarkDown.",
@@ -56,6 +57,7 @@ pixtral_vlm_conversion_options = HuggingFaceVlmOptions(
inference_framework=InferenceFramework.TRANSFORMERS_LlavaForConditionalGeneration,
)
vlm_conversion_options = pixtral_vlm_conversion_options
"""
"""
phi_vlm_conversion_options = HuggingFaceVlmOptions(
@@ -68,15 +70,13 @@ phi_vlm_conversion_options = HuggingFaceVlmOptions(
vlm_conversion_options = phi_vlm_conversion_options
"""
"""
pixtral_vlm_conversion_options = HuggingFaceVlmOptions(
repo_id="mlx-community/pixtral-12b-bf16",
prompt="Convert this full page to markdown. Do not miss any text and only output the bare MarkDown!",
prompt="Convert this page to markdown. Do not miss any text and only output the bare MarkDown!",
response_format=ResponseFormat.MARKDOWN,
inference_framework=InferenceFramework.MLX,
)
vlm_conversion_options = pixtral_vlm_conversion_options
"""
"""
qwen_vlm_conversion_options = HuggingFaceVlmOptions(