Update README.md
Browse files
README.md
CHANGED
@@ -36,25 +36,6 @@ transformers.AutoModelForCausalLM.from_pretrained
|
|
36 |
```
|
37 |
|
38 |
### Model Description
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
This is the model card of a 🤗 transformers model that has been pushed on the Hub.
|
43 |
-
Previous vision models have been 50/50 as the multimodel model actully requires a lot of memory and gpu and harddrive space to create;
|
44 |
-
the past versions have been attempts to Merge the capabilitys into the main mistral model whilst still retaining its mistral tag!
|
45 |
-
After reading many hugging face articles:
|
46 |
-
|
47 |
-
The BackBone Issue is the main cause of creating multi modals !:
|
48 |
-
|
49 |
-
with the advent of tiny models we are able to leverage the decoder abilitys as a single expert-ish... within the model :
|
50 |
-
by reducing the size to a fully trainined tiny model!
|
51 |
-
this will only produce decodings and not conversations so it needs to be smart and respond with defined answers: but in general it will produce captions: but as domain based it may be specialized in medical or art etc:
|
52 |
-
|
53 |
-
The main llm still needs to retain these models within hence the back bone method of instigating a VisionEncoderDecoder model: istead of a llava model which still need wrangling to work correctly without spoiling the original transformers installation:
|
54 |
-
Previous experiments proved that the mistral large model could be used as a decoder but the total model jumped to 13b so the when applying the tiny model it was only effected by the weight of the model 248M
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
This is an experiment in vision - the model has been created as a mistral/VisionEncoder/Decoder
|
59 |
|
60 |
Customized from:
|
@@ -80,6 +61,25 @@ Encoder:
|
|
80 |
- **Language(s) (NLP):** [English]
|
81 |
|
82 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
83 |
## How to Get Started with the Model
|
84 |
|
85 |
|
@@ -168,7 +168,12 @@ loss = model(pixel_values=pixel_values, labels=labels).loss
|
|
168 |
```
|
169 |
|
170 |
|
171 |
-
### Model Architecture
|
|
|
|
|
|
|
|
|
|
|
172 |
|
173 |
``` python
|
174 |
|
|
|
36 |
```
|
37 |
|
38 |
### Model Description
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
39 |
This is an experiment in vision - the model has been created as a mistral/VisionEncoder/Decoder
|
40 |
|
41 |
Customized from:
|
|
|
61 |
- **Language(s) (NLP):** [English]
|
62 |
|
63 |
|
64 |
+
## Summary
|
65 |
+
|
66 |
+
This is the model card of a 🤗 transformers model that has been pushed on the Hub.
|
67 |
+
Previous vision models have been 50/50 as the multimodel model actully requires a lot of memory and gpu and harddrive space to create;
|
68 |
+
the past versions have been attempts to Merge the capabilitys into the main mistral model whilst still retaining its mistral tag!
|
69 |
+
After reading many hugging face articles:
|
70 |
+
|
71 |
+
The BackBone Issue is the main cause of creating multi modals !:
|
72 |
+
|
73 |
+
with the advent of tiny models we are able to leverage the decoder abilitys as a single expert-ish... within the model :
|
74 |
+
by reducing the size to a fully trainined tiny model!
|
75 |
+
this will only produce decodings and not conversations so it needs to be smart and respond with defined answers: but in general it will produce captions: but as domain based it may be specialized in medical or art etc:
|
76 |
+
|
77 |
+
The main llm still needs to retain these models within hence the back bone method of instigating a VisionEncoderDecoder model: istead of a llava model which still need wrangling to work correctly without spoiling the original transformers installation:
|
78 |
+
Previous experiments proved that the mistral large model could be used as a decoder but the total model jumped to 13b so the when applying the tiny model it was only effected by the weight of the model 248M
|
79 |
+
|
80 |
+
|
81 |
+
|
82 |
+
|
83 |
## How to Get Started with the Model
|
84 |
|
85 |
|
|
|
168 |
```
|
169 |
|
170 |
|
171 |
+
### Model Architecture
|
172 |
+
|
173 |
+
|
174 |
+
Aha !!! Here is how you create such a model ::
|
175 |
+
|
176 |
+
|
177 |
|
178 |
``` python
|
179 |
|