Recent posts

#71

General Discussion and Questions / Re: Original vs. Duplicates

Last post by sybersitizen - March 23, 2025, 10:11:19 PM

"Originals" are the image you select for any match, including duplicates, visually similar images, sketch match, GPS search etc.
See Finding Duplicate Files for explanation of originals in this context.

Indeed, I had repeatedly missed a key sentence on that page, despite having visited it numerous times. My fault, no one else's.

But how can I be shown more than one file that is not red-lined?

Show where?

In the Result Window for Duplicates when I perform a search for duplicates. That's something I am still not understanding.

I like having IMatch help me with duplicates.

I like the red color coding for duplicates.

IMatch has identified a rather large number of duplicates because I originally had it ingest a collection of about 80,000 images that were a mix of things collected by three family members over decades. Many are binary duplicates and many of them are only slight variations. This will not be an issue going forward with new images.

I am not complaining about anything. I simply want to understand what exactly I'm being shown - and why - when I use the Result Window for Duplicates to consider cleaning up the duplicates. It is becoming a bit clearer.

#72

General Discussion and Questions / Gemma 3 12B on NVIDIA GeForce ...

Last post by jch2103 - March 23, 2025, 10:07:04 PM

I've run into some bumps using my RTX 4070 (12GB RAM) with IMatch/Gemma 3 12B. It seems that Gemma 3 12B requires every bit of the available memory on the GPU. There are times when I try to run AutoTagger with 12B after the computer's been used for a while and I get an error message (full log attached):

Code Select Expand

03.23 14:36:33+  282 [26F0C] 02  I> AIConnectorOpenAI: 1 HTTP Status Code: 404 'No models loaded. Please load a model in the developer page or use the `lms load` command.'
03.23 14:36:33+    0 [2684C] 01  W> AutoTagger: Aborting because of error
 'No models loaded. Please load a model in the developer page or use the `lms load` command.'  'V:\develop\IMatch5\src\IMEngine\IMEngineAIAutoTagger.cpp(369)'
03.23 14:36:33+  15 [29594] 01  W> UpdateQueue (AutoTagger): Service error 0 'No models loaded. Please load a model in the developer page or use the `lms load` command.' for file [143645]  'V:\develop\IMatch5\src\IMEngine\IMEngineUpdateQueueAutoTagger.cpp(233)'
If I switch to the LM Studio/Gemma 3 4B or OpenAI model, AutoTagger runs OK. I suspect the issue for 12B is lack of available GPU memory. 

If I reboot the computer and immediately run IMatch, I can run AutoTagger with Gemma 3 12B w/o problems. If my issue is indeed lack of GPU memory, is there a simpler way to clear its memory than a reboot?

Gemma 3 12B does seem to return more useful/accurate information than 4B, although OpenAI has expected advantages for things like more obscure landmarks. 

#73

General Discussion and Questions / Where is the Date coming from ...

Last post by kiwilink - March 23, 2025, 10:06:38 PM

I'm trying the new update (Ollama) and I'm noticing a date being returned.  Any idea where this date is from?  I ran Exiftool and this date doesn't show up.  I'm just curious.

Thanks

Kiwilink

#74

Solved Bug Reports (for next version) / Re: Renamer - copy file to a n...

Last post by Tveloso - March 23, 2025, 09:49:04 PM

Here is what the Preview shows here (the leading \\ in the UNC path is retained):

   

It's in the Edit Dialog for that Move step, where I see the \\ replaced with \:

   

...but that's not happening for me at runtime.

Also, the Preview in the ScreenShot from the OP has four columns (which translate to):

• Result
• Original Name
• Target Name
• Copy Name

...but mine has only three:

• Result
• Original Name
• Resulting Name

Is that a difference between a Copy and Move step?

#75

General Discussion and Questions / Re: Gamma3:4b (This model is n...

Last post by kiwilink - March 23, 2025, 09:40:10 PM

Updated Ollama and then all is okay!  Thank you!

#76

General Discussion and Questions / Re: Gamma3:4b (This model is n...

Last post by kiwilink - March 23, 2025, 09:28:20 PM

Mario

Thank you for the reply.  Yes, it returned 0.5.7.  Since I have an older version I assume I have to update Ollama first?

Thx, 
kiwilink

#77

General Discussion and Questions / Re: Gamma3:4b (This model is n...

Last post by Mario - March 23, 2025, 09:22:31 PM

Maybe your Ollama version is tool old?
When I run a

ollama -v

I get

ollama version is 0.6.2

Which version do you get? Ollama usually updates itself automatically.
When you right-click the Ollama icon in the task bar, does it show a pending update?

You can also uninstall Ollama and install LM Studio instead: LM Studio(Local AI)
It can run Gemma 3, works with IMatch and allows you to chat with Gemma3 or other models like you can chat with ChatGPT, Copilot or Gemini.

#78

General Discussion and Questions / Gamma3:4b (This model is not s...

Last post by kiwilink - March 23, 2025, 09:15:20 PM

I downloaded gamma3:4b and then got the message in the attachment.

Did I miss a step yja I was supposed to do before the download/install?

Thanks

Kiwilink

#79

General Discussion and Questions / Re: Original vs. Duplicates

Last post by Mario - March 23, 2025, 08:52:02 PM

"Originals" are the image you select for any match, including duplicates, visually similar images, sketch match, GPS search etc.
See Finding Duplicate Files for explanation of originals in this context.

But how can I be shown more than one file that is not red-lined?

Show where?

There are no "originals" in dupe search. if IMatch considers a file a dupe when ingesting it, it will be added to the category. What you consider the "original" is up to you.

If you don't want IMatch to indicate files it considers duplicates, change the settings under Edit > Preferences > Indexing. Disable the dupe search, switch it to 100% copies only, whatever works best for you.

If you find the red category color bar confusing, just disable the color-coding for the Duplicates category.

Files should be normally not long in Duplicates anyway.

IMatch reports some dupes after ingesting new files, you deal with them, e.g. selecting them and run a Search > for Duplicates/Copies and then decide which of the files (the original you selected for the search or the matches found) to keep. Then the files can be removed from the Duplicates category, making it clean for the next import.

#80

General Discussion and Questions / Re: Original vs. Duplicates

Last post by sybersitizen - March 23, 2025, 08:25:05 PM

In terms of doing a search for duplicates, and a Result Window opens showing "the originals" and one or more matching duplicates, the originals are simply the files that you have selected to perform the search.  So in performing the search, you have designated which files are the originals.

Thanks for the comments.

So IMatch doesn't try to distinguish 'originals' from 'duplicates' on its own - the 'original' is simply the photo I started the search with. If that was explained in the Help pages, I just failed to see it.

As Mario Described, the red line indicates that the file is in a Category that's configured with color coding (which the DUPLICATES Category is configured to do).

So in the scenario above, File B would have a red line, and File A would not.  This is not really directly related to any indication of which is the duplicate.  It simply indicates that the file has been added to a Category that does color coding (which File B has been).  As described in the Help Topic that Mario linked-to, you can set up your own categories to do color coding, and they would contribute their color to the thumbnail's color bar (just as the DUPLICATES category is doing).

All I use or need is the DUPLICATES category.

As I said, when I look at a group of duplicates, sometimes more than one are red-lined. Sometimes more than one are not red-lined. Sometimes the 'original' that I selected is red-lined and sometimes it is not.

I understand why I can be shown more than one file that is red-lined because those are considered duplicates, of which there can be many. But how can I be shown more than one file that is not red-lined? Shouldn't there only be one non-red-lined file in the group - the one that was indexed first? In an example I'm looking at right now, there are two non-red-lined files shown below the red-lined file that I selected for the search. They can't both be originals - but they are two different file types (JPEG and PSD) that were indexed in the same session. Does that explain it?