0.3428, cosine with 5e-5

README.md

@@ -7,161 +7,140 @@ tags:
 - sentence-similarity
 - feature-extraction
 - generated_from_trainer
-- dataset_size:2940
+- dataset_size:2632
 - loss:MultipleNegativesRankingLoss
 widget:
-- source_sentence: "Construct:  Identify a unit of area.\n\nQuestion:  James has answered\
-    \ a question on the area of a trapezium and got an answer of  54 .\n\nBehind the\
-    \ star he has written the units that he used.\n\n(\n54 , * \n)\n\nWhich of the\
-    \ following units could be correct?\n\nOptions:\nA. m\nB. cm\nC. km^3\nD. mm^2\n\
-    \nAnswer: cm"
-  sentences:
-  - Identify a unit of area
-  - "James has answered a question on the area of a trapezium and got an answer of\
-    \  54 .\n\nBehind the star he has written the units that he used.\n\n(\n54 , *\
-    \ \n)\n\nWhich of the following units could be correct?"
-  - Does not know units of area should be squared
-- source_sentence: 'Construct:  Can calculate missing coordinates in a geometric shape.
-
-
-    Question:  Here are  3  vertices of a rectangle:
+- source_sentence: 'Construct:  Identify the line of symmetry of curves in the form
+    (x + a)(x + b).
 
-    (0,3),(4,-1),(2,-3),
 
-
-    What are the coordinates of the  4^th   vertex?
+    Question:  This is the graph of  y=(x+4)(x-2)  W What is the equation of the line
+    of symmetry?
 
 
     Options:
 
-    A. (0,1)
+    A. y=-8
 
-    B. (4,1)
+    B. x=--4
 
-    C. (-2,1)
+    C. y=-1
 
-    D. (-2,-3)
+    D. x = -1
 
 
-    Answer: (4,1)'
+    Answer: y=-1'
   sentences:
-  - 'Here are  3  vertices of a rectangle:
-
-    (0,3),(4,-1),(2,-3),
+  - This is the graph of  y=(x+4)(x-2)  W What is the equation of the line of symmetry?
+  - Identify the line of symmetry of curves in the form (x + a)(x + b)
+  - Confuses the equations of vertical and horizontal lines
+- source_sentence: 'Construct:  Calculate percentage of an amount mentally where the
+    percent is a decimal.
 
 
-    What are the coordinates of the  4^th   vertex?'
-  - Can calculate missing coordinates in a geometric shape
-  - Believes the sides of a rectangle must be horizontal and vertical, not diagonal
-- source_sentence: 'Construct:  Reflect an object or a point in a diagonal line with
-    the line of reflection drawn, where the line of reflection lies on the edge or
-    outside the object.
-
-
-    Question:  The triangle is reflected in the dashed line
-
-    What are the new coordinates of point  P  ? Four quadrant, coordinate grid with
-    the reflection line y=-x drawn and a triangle. The triangle has coordinates: (-2,3)
-    (-2,6) and (0,5). The point P is the coordinate (0,5).
+    Question:  What is  5.5 %  of  400  ?
 
 
     Options:
 
-    A. (-5,0)
+    A. 22
 
-    B. (-4,-1)
+    B. 11
 
-    C. (-2,2)
+    C. 5.5
 
-    D. (-4,5)
+    D. 220
 
 
-    Answer: (-4,5)'
+    Answer: 220'
   sentences:
-  - Reflects horizontally or vertically instead of across a diagonal line of symmetry
-  - Reflect an object or a point in a diagonal line with the line of reflection drawn,
-    where the line of reflection lies on the edge or outside the object
-  - 'The triangle is reflected in the dashed line
+  - Believes that dividing by 10 gives 1%
+  - What is  5.5 %  of  400  ?
+  - Calculate percentage of an amount mentally where the percent is a decimal
+- source_sentence: 'Construct:  Convert from weeks to hours.
 
-    What are the new coordinates of point  P  ? Four quadrant, coordinate grid with
-    the reflection line y=-x drawn and a triangle. The triangle has coordinates: (-2,3)
-    (-2,6) and (0,5). The point P is the coordinate (0,5)'
-- source_sentence: 'Construct:  Rearrange formulae to change the subject where the
-    subject appears once and two steps are needed.
 
+    Question:  Which one of the following calculations would work out the number of
+    hours in  5  weeks?
 
-    Question:  Rearrange the following equation to make  r  the subject
 
-    [
+    Options:
 
-    p=3 r-5
+    A. 5 x 24
 
-    ].
+    B. 5 x 7 x 12
 
+    C. 5 x 5 x 24
 
-    Options:
+    D. 5 x 7 x 24
 
-    A. r=(p+5 / 3)
 
-    B. r=p+5 / 3
+    Answer: 5 x 5 x 24'
+  sentences:
+  - Which one of the following calculations would work out the number of hours in  5  weeks?
+  - Convert from weeks to hours
+  - Thinks there are 5 days in a week
+- source_sentence: 'Construct:  Expand two brackets with linear terms in the form
+    (ax + b)².
 
-    C. r=(p-5 / 3)
 
-    D. r=p-5 / 3
+    Question:  Expand and simplify:
 
+    (
 
-    Answer: r=p+5 / 3'
-  sentences:
-  - Rearrange formulae to change the subject where the subject appears once and two
-    steps are needed
-  - 'Rearrange the following equation to make  r  the subject
+    (3 y-4)^2
 
-    [
+    ).
 
-    p=3 r-5
 
-    ]'
-  - Misunderstands order of operations in algebraic expressions
-- source_sentence: 'Construct:  Solve length problems involving similar shapes where
-    the missing length is the shorter one.
+    Options:
 
+    A. 9 y^2-16
 
-    Question:  These two shapes are similar.
+    B. 9 y^2+16
 
+    C. 9 y^2-24 y+16
 
-    What is the value of  y  ? The diagram is of two trapeziums. The base of each
-    trapezium is parallel with its top side.
+    D. (3 y-4)(3 y-4)
 
-    The smaller trapezium has base 60cm and top labelled y.
 
-    The larger trapezium has base 150cm and top 70cm.
+    Answer: 9 y^2+16'
+  sentences:
+  - Believes they only need to multiply the first and last pairs of terms when expanding
+    double brackets
+  - Expand two brackets with linear terms in the form (ax + b)²
+  - 'Expand and simplify:
 
+    (
 
-    Options:
+    (3 y-4)^2
 
-    A. 28 cm
+    )'
+- source_sentence: 'Construct:  Write a given translation in vector notation.
 
-    B. 20 cm
 
-    C. 25 cm
+    Question:  Three left and five down in vector notation is...  triangle  bigcirc  What
+    should replace the  triangle  ?
 
-    D. 30 cm
 
+    Options:
 
-    Answer: 30 cm'
-  sentences:
-  - 'These two shapes are similar.
+    A. 3
 
+    B. -3
 
-    What is the value of  y  ? The diagram is of two trapeziums. The base of each
-    trapezium is parallel with its top side.
+    C. 5
 
-    The smaller trapezium has base 60cm and top labelled y.
+    D. -5
 
-    The larger trapezium has base 150cm and top 70cm.'
-  - Estimated when not appropriate
-  - Solve length problems involving similar shapes where the missing length is the
-    shorter one
+
+    Answer: -5'
+  sentences:
+  - Write a given translation in vector notation
+  - Thinks the movement of the top number in a column vector is vertical, and the
+    bottom number is horizontal rather than the other way round
+  - Three left and five down in vector notation is...  triangle  bigcirc  What should
+    replace the  triangle  ?
 ---
 
 # SentenceTransformer based on Alibaba-NLP/gte-base-en-v1.5
@@ -214,9 +193,9 @@ from sentence_transformers import SentenceTransformer
 model = SentenceTransformer("Gurveer05/gte-base-eedi-2024")
 # Run inference
 sentences = [
-    'Construct:  Solve length problems involving similar shapes where the missing length is the shorter one.\n\nQuestion:  These two shapes are similar.\n\nWhat is the value of  y  ? The diagram is of two trapeziums. The base of each trapezium is parallel with its top side.\nThe smaller trapezium has base 60cm and top labelled y.\nThe larger trapezium has base 150cm and top 70cm.\n\nOptions:\nA. 28 cm\nB. 20 cm\nC. 25 cm\nD. 30 cm\n\nAnswer: 30 cm',
-    'Estimated when not appropriate',
-    'Solve length problems involving similar shapes where the missing length is the shorter one',
+    'Construct:  Write a given translation in vector notation.\n\nQuestion:  Three left and five down in vector notation is...  triangle  bigcirc  What should replace the  triangle  ?\n\nOptions:\nA. 3\nB. -3\nC. 5\nD. -5\n\nAnswer: -5',
+    'Thinks the movement of the top number in a column vector is vertical, and the bottom number is horizontal rather than the other way round',
+    'Write a given translation in vector notation',
 ]
 embeddings = model.encode(sentences)
 print(embeddings.shape)
@@ -271,19 +250,19 @@ You can finetune this model on your own dataset.
 #### csv
 
 * Dataset: csv
-* Size: 2,940 training samples
+* Size: 2,632 training samples
 * Columns: <code>qa_pair_text</code>, <code>MisconceptionName</code>, <code>ConstructName</code>, and <code>QuestionText</code>
 * Approximate statistics based on the first 1000 samples:
   |         | qa_pair_text                                                                        | MisconceptionName                                                                 | ConstructName                                                                     | QuestionText                                                                       |
   |:--------|:------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|
   | type    | string                                                                              | string                                                                            | string                                                                            | string                                                                             |
-  | details | <ul><li>min: 35 tokens</li><li>mean: 94.91 tokens</li><li>max: 903 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 14.61 tokens</li><li>max: 39 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 14.39 tokens</li><li>max: 41 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 38.43 tokens</li><li>max: 278 tokens</li></ul> |
+  | details | <ul><li>min: 35 tokens</li><li>mean: 92.39 tokens</li><li>max: 507 tokens</li></ul> | <ul><li>min: 4 tokens</li><li>mean: 15.37 tokens</li><li>max: 39 tokens</li></ul> | <ul><li>min: 4 tokens</li><li>mean: 14.01 tokens</li><li>max: 42 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 38.59 tokens</li><li>max: 274 tokens</li></ul> |
 * Samples:
-  | qa_pair_text                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         | MisconceptionName                                                                                                                | ConstructName                                                                               | QuestionText                                                                                                                                                                                                                                                  |
-  |:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
-  | <code>Construct:  Work out the size of each interior angle in a regular polygon.<br><br>Question:  The octagon below is regular. Find angle  x . A regular octagon (8 sided polygon). Each side is marked with a single dash to show that all sides are equal in length. One of the interior angles of the octagon is labelled with the letter "x".<br><br>Options:<br>A. 22.5°<br>B. 120°<br>C. 135°<br>D. 45°<br><br>Answer: 22.5°</code>                                                                                                          | <code>Thinks the interior angles of any polygon add up to 180 degrees</code>                                                     | <code>Work out the size of each interior angle in a regular polygon</code>                  | <code>The octagon below is regular. Find angle  x . A regular octagon (8 sided polygon). Each side is marked with a single dash to show that all sides are equal in length. One of the interior angles of the octagon is labelled with the letter "x".</code> |
-  | <code>Construct:  Manipulate both equations so that they are in a form to be added or subtracted.<br><br>Question:  Sandeep wants to eliminate  l  from these simultaneous equations.<br>(<br>2 l+3 m=13 <br>3 l+4 m=17<br>)<br><br>First he will transform both of them to get them in an appropriate form.<br>Which of the following is a correct transformation?<br><br>Options:<br>A. 6 l+9 m=26  6 l+8 m=51<br>B. 6 l+9 m=13  6 l+8 m=17<br>C. 6 l+3 m=13  6 l+4 m=17<br>D. 6 l+9 m=39  6 l+8 m=34<br><br>Answer: 6 l+9 m=13  6 l+8 m=17</code> | <code>When solving simultaneous equations, thinks you only transform the algebraic terms, rather than the whole equation.</code> | <code>Manipulate both equations so that they are in a form to be added or subtracted</code> | <code>Sandeep wants to eliminate  l  from these simultaneous equations.<br>(<br>2 l+3 m=13 <br>3 l+4 m=17<br>)<br><br>First he will transform both of them to get them in an appropriate form.<br>Which of the following is a correct transformation?</code>  |
-  | <code>Construct:  Express one quantity as a percentage of another mentally.<br><br>Question:  What is  8  out of  40  as a percentage?<br><br>Options:<br>A. 8.4 %<br>B. 20 %<br>C. 16 %<br>D. 24 %<br><br>Answer: 16 %</code>                                                                                                                                                                                                                                                                                                                       | <code>Thinks they double the numerator to turn a fraction into a percentage</code>                                               | <code>Express one quantity as a percentage of another mentally</code>                       | <code>What is  8  out of  40  as a percentage?</code>                                                                                                                                                                                                         |
+  | qa_pair_text                                                                                                                                                                                                                                                                                                                                                                                                 | MisconceptionName                                                                            | ConstructName                                                                                               | QuestionText                                                                                                                                             |
+  |:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------|
+  | <code>Construct:  Given the name of a polygon, draw or identify a corresponding diagram.<br><br>Question:  Which of the following shapes is a hexagon?<br><br>Options:<br>A. An irregular, five sided shape<br>B. A regular, five sided shape<br>C. A regular, eight sided shape<br>D. None of these<br><br>Answer: A regular, eight sided shape</code>                                                      | <code>Confuses octagon and hexagon</code>                                                    | <code>Given the name of a polygon, draw or identify a corresponding diagram</code>                          | <code>Which of the following shapes is a hexagon?</code>                                                                                                 |
+  | <code>Construct:  Rearrange formulae to change the subject where the subject appears once and one step is needed.<br><br>Question:  Step  2 <br><br>Which is the correct next step rearrangement of the following formula, making  u  the subject?<br>[<br>7 t-13=u(5+2 t)<br>].<br><br>Options:<br>A. u=5 t-8<br>B. u=(7 t-13 / 5+2 t)<br>C. u=(5+2 t / 7 t-13)<br>D. u=9 t-8<br><br>Answer: u=9 t-8</code> | <code>Misunderstands the process of making a variable the subject of an equation</code>      | <code>Rearrange formulae to change the subject where the subject appears once and one step is needed</code> | <code>Step  2 <br><br>Which is the correct next step rearrangement of the following formula, making  u  the subject?<br>[<br>7 t-13=u(5+2 t)<br>]</code> |
+  | <code>Construct:  Divide proper fractions in the form: Fraction ÷ Integer.<br><br>Question:  Calculate<br>(<br>(2 / 5) / 3<br>).<br><br>Options:<br>A. (6 / 5)<br>B. (15 / 2)<br>C. (2 / 15)<br>D. (6 / 15)<br><br>Answer: (6 / 5)</code>                                                                                                                                                                    | <code>when dividing a fraction by an integer, multiplies the numerator by the integer</code> | <code>Divide proper fractions in the form: Fraction ÷ Integer</code>                                        | <code>Calculate<br>(<br>(2 / 5) / 3<br>)</code>                                                                                                          |
 * Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
   ```json
   {
@@ -297,19 +276,19 @@ You can finetune this model on your own dataset.
 #### csv
 
 * Dataset: csv
-* Size: 1,430 evaluation samples
+* Size: 1,738 evaluation samples
 * Columns: <code>qa_pair_text</code>, <code>MisconceptionName</code>, <code>ConstructName</code>, and <code>QuestionText</code>
 * Approximate statistics based on the first 1000 samples:
   |         | qa_pair_text                                                                        | MisconceptionName                                                                 | ConstructName                                                                     | QuestionText                                                                       |
   |:--------|:------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|
   | type    | string                                                                              | string                                                                            | string                                                                            | string                                                                             |
-  | details | <ul><li>min: 35 tokens</li><li>mean: 95.63 tokens</li><li>max: 903 tokens</li></ul> | <ul><li>min: 4 tokens</li><li>mean: 14.59 tokens</li><li>max: 40 tokens</li></ul> | <ul><li>min: 4 tokens</li><li>mean: 14.26 tokens</li><li>max: 41 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 39.45 tokens</li><li>max: 212 tokens</li></ul> |
+  | details | <ul><li>min: 35 tokens</li><li>mean: 95.35 tokens</li><li>max: 903 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 14.36 tokens</li><li>max: 40 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 14.44 tokens</li><li>max: 40 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 37.54 tokens</li><li>max: 212 tokens</li></ul> |
 * Samples:
-  | qa_pair_text                                                                                                                                                                                                                                                                                                                                                                                            | MisconceptionName                                                                                        | ConstructName                                                                        | QuestionText                                                                                                                                                |
-  |:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------------------------------------------------------------------------------|
-  | <code>Construct:  Calculate the mean as an average from a list of data.<br><br>Question:  Without using a calculator, what is the mean of the following numbers?<br><br>[<br>2,2,3,3,2<br>].<br><br>Options:<br>A. 2<br>B. A little more than  2<br>C. 3<br>D. A little less than  2<br><br>Answer: 2</code>                                                                                            | <code>When asked for the mean of a list of data, gives the mode</code>                                   | <code>Calculate the mean as an average from a list of data</code>                    | <code>Without using a calculator, what is the mean of the following numbers?<br><br>[<br>2,2,3,3,2<br>]</code>                                              |
-  | <code>Construct:  Solve problems that combine addition or subtraction with multiplication.<br><br>Question:  Sara has  £ 2.79 .<br>Rhys has  3  times as much as Sara.<br>How much do they have altogether?<br>Which calculation would not find the answer?<br><br>Options:<br>A. £ 2.79 x 4<br>B. (£ 2.79 x 3)+£ 2.79<br>C. £ 2.80 x 4-£ 0.04<br>D. £ 2.79 x 3<br><br>Answer: £ 2.80 x 4-£ 0.04</code> | <code>Does not think that the compensation strategy for the multiplication of 2 numbers is valid.</code> | <code>Solve problems that combine addition or subtraction with multiplication</code> | <code>Sara has  £ 2.79 .<br>Rhys has  3  times as much as Sara.<br>How much do they have altogether?<br>Which calculation would not find the answer?</code> |
-  | <code>Construct:  Recognise and name an identity.<br><br>Question:  Which of the following is the most appropriate name for this:<br>p-3 equiv-(3-p).<br><br>Options:<br>A. Equation<br>B. Expression<br>C. Formula<br>D. Identity<br><br>Answer: Equation</code>                                                                                                                                       | <code>Confuses an equation with an identity</code>                                                       | <code>Recognise and name an identity</code>                                          | <code>Which of the following is the most appropriate name for this:<br>p-3 equiv-(3-p)</code>                                                               |
+  | qa_pair_text                                                                                                                                                                                                                                                            | MisconceptionName                                                                  | ConstructName                                                                                    | QuestionText                                                                         |
+  |:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|
+  | <code>Construct:  Add proper fractions with different denominators which do not share a common factor.<br><br>Question:  (<br>(1 / 3)+(1 / 10)=(* / 30)<br>)<br>What should replace the star?<br><br>Options:<br>A. 13<br>B. 2<br>C. 4<br>D. 11<br><br>Answer: 2</code> | <code>Does not find a common denominator when adding/subtracting fractions</code>  | <code>Add proper fractions with different denominators which do not share a common factor</code> | <code>(<br>(1 / 3)+(1 / 10)=(* / 30)<br>)<br>What should replace the star?</code>    |
+  | <code>Construct:  Round decimals to the nearest whole number.<br><br>Question:  What is  9 0 . 6 8 2  rounded to the nearest integer?<br><br>Options:<br>A. 90<br>B. 90.7<br>C. 91<br>D. 92<br><br>Answer: 90.7</code>                                                  | <code>Rounds to the wrong degree of accuracy (rounds too much)</code>              | <code>Round decimals to the nearest whole number</code>                                          | <code>What is  9 0 . 6 8 2  rounded to the nearest integer?</code>                   |
+  | <code>Construct:  Subtract fractions involving mixed number fractions with different denominators.<br><br>Question:  (<br>2 (2 / 5)-(1 / 15)=2 (1 / *)<br>)<br>What should replace the star?<br><br>Options:<br>A. 10<br>B. 15<br>C. 5<br>D. 3<br><br>Answer: 10</code> | <code>When subtracting fractions, subtracts the numerators and denominators</code> | <code>Subtract fractions involving mixed number fractions with different denominators</code>     | <code>(<br>2 (2 / 5)-(1 / 15)=2 (1 / *)<br>)<br>What should replace the star?</code> |
 * Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
   ```json
   {
@@ -450,24 +429,22 @@ You can finetune this model on your own dataset.
 </details>
 
 ### Training Logs
-| Epoch   | Step    | Training Loss | loss       |
-|:-------:|:-------:|:-------------:|:----------:|
-| 0.25    | 46      | 4.1456        | -          |
-| 0.5     | 92      | 1.3677        | 0.8336     |
-| 0.75    | 138     | 0.7569        | -          |
-| 1.0     | 184     | 0.651         | 0.4735     |
-| 1.25    | 230     | 0.587         | -          |
-| 1.5     | 276     | 0.4073        | 0.4675     |
-| 1.75    | 322     | 0.3533        | -          |
-| 2.0     | 368     | 0.348         | 0.4002     |
-| 2.25    | 414     | 0.308         | -          |
-| 2.5     | 460     | 0.2247        | 0.3935     |
-| 2.75    | 506     | 0.194         | -          |
-| **3.0** | **552** | **0.1819**    | **0.3638** |
-| 3.25    | 598     | 0.1758        | -          |
-| 3.5     | 644     | 0.1247        | 0.4158     |
-| 3.75    | 690     | 0.1035        | -          |
-| 4.0     | 736     | 0.1155        | 0.4073     |
+| Epoch      | Step    | Training Loss | loss       |
+|:----------:|:-------:|:-------------:|:----------:|
+| 0.2545     | 42      | 4.1578        | -          |
+| 0.5030     | 83      | -             | 1.0737     |
+| 0.5091     | 84      | 1.4123        | -          |
+| 0.7636     | 126     | 0.7791        | -          |
+| 1.0061     | 166     | -             | 0.7048     |
+| 1.0182     | 168     | 0.5538        | -          |
+| 1.2727     | 210     | 0.5208        | -          |
+| **1.5091** | **249** | **-**         | **0.5806** |
+| 1.5273     | 252     | 0.428         | -          |
+| 1.7818     | 294     | 0.3175        | -          |
+| 2.0121     | 332     | -             | 0.6534     |
+| 2.0364     | 336     | 0.2301        | -          |
+| 2.2909     | 378     | 0.2247        | -          |
+| 2.5152     | 415     | -             | 0.6442     |
 
 * The bold row denotes the saved checkpoint.
 

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0e86404d491b64c16bd9f3112a8b6abea6c8ae1efb0aefaaf5cb170bf4b5dff0
+oid sha256:bfd2d15951004574fa79cd6f336c5bafaea48639523392e277aaa60efd5f8ca8
 size 547119128