Skip to content

Commit

Permalink
Improve qwen vl impl (#2943)
Browse files Browse the repository at this point in the history
* feat: refactor model, improve startup and re enable tests

* fix: improve multimodal rotary embed caching

* fix: limit vision flop calc to qwen2 vl models and update config typing

* fix: include clippy lint

* feat: refactor position ids in warmup and bump tests

* fix: prefer default dtype

* fix: enable all cuda graphs and bump snapshots

* fix: adjust rotaty init path

* fix: simplify get position ids and remove usused vision config

* fix: update position ids so first dim is batch, simplify rotary and bump vlm default token limit

* fix: improve position id init during cuda warmup for mrope and simplfy rotary forward

* fix: check existance before accessing rope type in cuda warmup

* fix: check key before access

* fix: improve mrope check in cuda graph warmup

* fix: remove check for default rope type

* fix: add more test and improve model generation

* fix: improve and simplify get_cos_sin, refactors and cleanup  get_position_ids

* fix: adjust signatures with types
  • Loading branch information
drbh authored Feb 4, 2025
1 parent dd2bd5f commit c1cf36c
Show file tree
Hide file tree
Showing 15 changed files with 850 additions and 309 deletions.

Large diffs are not rendered by default.

Original file line number Diff line number Diff line change
Expand Up @@ -7,67 +7,67 @@
"seed": 0,
"tokens": [
{
"id": 1939,
"logprob": -2.2460938,
"id": 5267,
"logprob": -1.1464844,
"special": false,
"text": "?\n\n"
"text": "?\n"
},
{
"id": 33464,
"logprob": 0.0,
"logprob": -0.83203125,
"special": false,
"text": "Deep"
},
{
"id": 20909,
"logprob": -0.48608398,
"logprob": -0.5625,
"special": false,
"text": " Learning"
},
{
"id": 4102,
"logprob": -2.265625,
"id": 320,
"logprob": -2.1464844,
"special": false,
"text": " "
"text": " ("
},
{
"id": 285,
"id": 16524,
"logprob": 0.0,
"special": false,
"text": "is"
"text": "DL"
},
{
"id": 458,
"logprob": -0.6328125,
"id": 701,
"logprob": -2.2089844,
"special": false,
"text": " an"
"text": "),"
},
{
"id": 20443,
"logprob": -0.1796875,
"id": 476,
"logprob": -0.27368164,
"special": false,
"text": " artificial"
"text": " or"
},
{
"id": 11229,
"logprob": 0.0,
"id": 20443,
"logprob": -0.09442139,
"special": false,
"text": " intelligence"
"text": " artificial"
},
{
"id": 320,
"logprob": -0.37695312,
"id": 29728,
"logprob": 0.0,
"special": false,
"text": " ("
"text": " neural"
},
{
"id": 15469,
"id": 14155,
"logprob": 0.0,
"special": false,
"text": "AI"
"text": " networks"
}
],
"top_tokens": null
},
"generated_text": "What is deep learning?\n\nDeep Learning is an artificial intelligence (AI"
"generated_text": "What is deep learning?\nDeep Learning (DL), or artificial neural networks"
}
Original file line number Diff line number Diff line change
Expand Up @@ -9,61 +9,61 @@
"tokens": [
{
"id": 18183,
"logprob": -1.4912109,
"logprob": -1.5195312,
"special": false,
"text": " Deep"
},
{
"id": 6832,
"logprob": -0.075683594,
"logprob": -0.06817627,
"special": false,
"text": " learning"
},
{
"id": 374,
"logprob": -0.12408447,
"logprob": -0.13122559,
"special": false,
"text": " is"
},
{
"id": 264,
"logprob": -0.12768555,
"logprob": -0.13415527,
"special": false,
"text": " a"
},
{
"id": 25993,
"logprob": -0.82128906,
"logprob": -0.87353516,
"special": false,
"text": " subset"
},
{
"id": 315,
"logprob": -0.0012636185,
"logprob": -0.0011396408,
"special": false,
"text": " of"
},
{
"id": 5662,
"logprob": -0.12878418,
"logprob": -0.16442871,
"special": false,
"text": " machine"
},
{
"id": 6832,
"logprob": -0.0015888214,
"logprob": -0.0026416779,
"special": false,
"text": " learning"
},
{
"id": 429,
"logprob": -0.49194336,
"logprob": -0.48754883,
"special": false,
"text": " that"
},
{
"id": 5711,
"logprob": -1.2626953,
"logprob": -1.2294922,
"special": false,
"text": " uses"
}
Expand All @@ -82,61 +82,61 @@
"tokens": [
{
"id": 18183,
"logprob": -1.4912109,
"logprob": -1.5195312,
"special": false,
"text": " Deep"
},
{
"id": 6832,
"logprob": -0.075683594,
"logprob": -0.06817627,
"special": false,
"text": " learning"
},
{
"id": 374,
"logprob": -0.12408447,
"logprob": -0.13122559,
"special": false,
"text": " is"
},
{
"id": 264,
"logprob": -0.12768555,
"logprob": -0.13415527,
"special": false,
"text": " a"
},
{
"id": 25993,
"logprob": -0.82128906,
"logprob": -0.87353516,
"special": false,
"text": " subset"
},
{
"id": 315,
"logprob": -0.0012636185,
"logprob": -0.0011396408,
"special": false,
"text": " of"
},
{
"id": 5662,
"logprob": -0.12878418,
"logprob": -0.16442871,
"special": false,
"text": " machine"
},
{
"id": 6832,
"logprob": -0.0015888214,
"logprob": -0.0026416779,
"special": false,
"text": " learning"
},
{
"id": 429,
"logprob": -0.49194336,
"logprob": -0.48754883,
"special": false,
"text": " that"
},
{
"id": 5711,
"logprob": -1.2626953,
"logprob": -1.2294922,
"special": false,
"text": " uses"
}
Expand All @@ -155,61 +155,61 @@
"tokens": [
{
"id": 18183,
"logprob": -1.4912109,
"logprob": -1.5195312,
"special": false,
"text": " Deep"
},
{
"id": 6832,
"logprob": -0.075683594,
"logprob": -0.06817627,
"special": false,
"text": " learning"
},
{
"id": 374,
"logprob": -0.12408447,
"logprob": -0.13122559,
"special": false,
"text": " is"
},
{
"id": 264,
"logprob": -0.12768555,
"logprob": -0.13415527,
"special": false,
"text": " a"
},
{
"id": 25993,
"logprob": -0.82128906,
"logprob": -0.87353516,
"special": false,
"text": " subset"
},
{
"id": 315,
"logprob": -0.0012636185,
"logprob": -0.0011396408,
"special": false,
"text": " of"
},
{
"id": 5662,
"logprob": -0.12878418,
"logprob": -0.16442871,
"special": false,
"text": " machine"
},
{
"id": 6832,
"logprob": -0.0015888214,
"logprob": -0.0026416779,
"special": false,
"text": " learning"
},
{
"id": 429,
"logprob": -0.49194336,
"logprob": -0.48754883,
"special": false,
"text": " that"
},
{
"id": 5711,
"logprob": -1.2626953,
"logprob": -1.2294922,
"special": false,
"text": " uses"
}
Expand All @@ -228,61 +228,61 @@
"tokens": [
{
"id": 18183,
"logprob": -1.4912109,
"logprob": -1.5195312,
"special": false,
"text": " Deep"
},
{
"id": 6832,
"logprob": -0.075683594,
"logprob": -0.06817627,
"special": false,
"text": " learning"
},
{
"id": 374,
"logprob": -0.12408447,
"logprob": -0.13122559,
"special": false,
"text": " is"
},
{
"id": 264,
"logprob": -0.12768555,
"logprob": -0.13415527,
"special": false,
"text": " a"
},
{
"id": 25993,
"logprob": -0.82128906,
"logprob": -0.87353516,
"special": false,
"text": " subset"
},
{
"id": 315,
"logprob": -0.0012636185,
"logprob": -0.0011396408,
"special": false,
"text": " of"
},
{
"id": 5662,
"logprob": -0.12878418,
"logprob": -0.16442871,
"special": false,
"text": " machine"
},
{
"id": 6832,
"logprob": -0.0015888214,
"logprob": -0.0026416779,
"special": false,
"text": " learning"
},
{
"id": 429,
"logprob": -0.49194336,
"logprob": -0.48754883,
"special": false,
"text": " that"
},
{
"id": 5711,
"logprob": -1.2626953,
"logprob": -1.2294922,
"special": false,
"text": " uses"
}
Expand Down
Loading

0 comments on commit c1cf36c

Please sign in to comment.