* ✅ Simplify tests for request_forms. PR [#13184](https://github.com/fastapi/fastapi/pull/13184) by [@alejsdev](https://github.com/alejsdev).
* ✅ Simplify tests for path_query_params. PR [#13181](https://github.com/fastapi/fastapi/pull/13181) by [@alejsdev](https://github.com/alejsdev).

        `compression` argument.

*   Keras:

    *   `tf.keras.layers.Conv` now includes a public `convolution_op` method.
        This method can be used to simplify the implementation of Conv
        subclasses. There are two primary ways to use this new method. The first
        is to use the method directly in your own `call` method: `python class

$table-cell-vertical-align;\n  border-color: var(--#{$prefix}table-border-color);\n\n  // Target th & td\n  // We need the child combinator to prevent styles leaking to nested tables which doesn't have a `.table` class.\n  // We use the universal selectors here to simplify the selector (else we would need 6 different selectors).\n  // Another advantage is that this generates less code and makes the selector less specific making it easier to override.\n  // stylelint-disable-next-line selector-max-universal\n  > :not(caption)...

.custom-file-label::after {\n      @include border-right-radius(0);\n    }\n  }\n}\n\n\n// Prepend and append\n//\n// While it requires one extra layer of HTML for each, dedicated prepend and\n// append elements allow us to 1) be less clever, 2) simplify our selectors, and\n// 3) support HTML5 form validation.\n\n.input-group-prepend,\n.input-group-append {\n  display: flex;\n\n  // Ensure buttons are always above inputs for more visually pleasing borders.\n  // This isn't needed for `.input-group-text`...

Step 2 and then // discards it in Step 6. d := bLimbs[i] c1 := addMulVVW(T[i:n+i], aLimbs, d) // Step 6 is replaced by shifting the virtual window we operate // over: T of the algorithm is T[i:] for us. That means that T1 in // Step 2 (T mod 2^_W) is simply T[i]. k0 in Step 3 is our m0inv. Y := T[i] * m.m0inv // Step 4 and 5 add Y × m to T, which as mentioned above is stored // at T[i:]. The two carries (from a × d and Y × m) are added up in // the next word T[n+i], and the carry bit from that addition...

Step 2 and then // discards it in Step 6. d := bLimbs[i] c1 := addMulVVW(T[i:n+i], aLimbs, d) // Step 6 is replaced by shifting the virtual window we operate // over: T of the algorithm is T[i:] for us. That means that T1 in // Step 2 (T mod 2^_W) is simply T[i]. k0 in Step 3 is our m0inv. Y := T[i] * m.m0inv // Step 4 and 5 add Y × m to T, which as mentioned above is stored // at T[i:]. The two carries (from a × d and Y × m) are added up in // the next word T[n+i], and the carry bit from that addition...