Add files using upload-large-folder tool

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/helpers/__init__.py

@@ -0,0 +1,6 @@
+"""Functions for parsing Links"""
+
+__all__ = ("parseLinkDestination", "parseLinkLabel", "parseLinkTitle")
+from .parse_link_destination import parseLinkDestination
+from .parse_link_label import parseLinkLabel
+from .parse_link_title import parseLinkTitle

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/helpers/parse_link_destination.py

@@ -0,0 +1,83 @@
+"""
+Parse link destination
+"""
+
+from ..common.utils import charCodeAt, unescapeAll
+
+
+class _Result:
+    __slots__ = ("ok", "pos", "str")
+
+    def __init__(self) -> None:
+        self.ok = False
+        self.pos = 0
+        self.str = ""
+
+
+def parseLinkDestination(string: str, pos: int, maximum: int) -> _Result:
+    start = pos
+    result = _Result()
+
+    if charCodeAt(string, pos) == 0x3C:  # /* < */
+        pos += 1
+        while pos < maximum:
+            code = charCodeAt(string, pos)
+            if code == 0x0A:  # /* \n */)
+                return result
+            if code == 0x3C:  # / * < * /
+                return result
+            if code == 0x3E:  # /* > */) {
+                result.pos = pos + 1
+                result.str = unescapeAll(string[start + 1 : pos])
+                result.ok = True
+                return result
+
+            if code == 0x5C and pos + 1 < maximum:  # \
+                pos += 2
+                continue
+
+            pos += 1
+
+        # no closing '>'
+        return result
+
+    # this should be ... } else { ... branch
+
+    level = 0
+    while pos < maximum:
+        code = charCodeAt(string, pos)
+
+        if code is None or code == 0x20:
+            break
+
+        # ascii control characters
+        if code < 0x20 or code == 0x7F:
+            break
+
+        if code == 0x5C and pos + 1 < maximum:
+            if charCodeAt(string, pos + 1) == 0x20:
+                break
+            pos += 2
+            continue
+
+        if code == 0x28:  # /* ( */)
+            level += 1
+            if level > 32:
+                return result
+
+        if code == 0x29:  # /* ) */)
+            if level == 0:
+                break
+            level -= 1
+
+        pos += 1
+
+    if start == pos:
+        return result
+    if level != 0:
+        return result
+
+    result.str = unescapeAll(string[start:pos])
+    result.pos = pos
+    result.ok = True
+    return result

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/helpers/parse_link_label.py

@@ -0,0 +1,44 @@
+"""
+Parse link label
+
+this function assumes that first character ("[") already matches
+returns the end of the label
+
+"""
+
+from markdown_it.rules_inline import StateInline
+
+
+def parseLinkLabel(state: StateInline, start: int, disableNested: bool = False) -> int:
+    labelEnd = -1
+    oldPos = state.pos
+    found = False
+
+    state.pos = start + 1
+    level = 1
+
+    while state.pos < state.posMax:
+        marker = state.src[state.pos]
+        if marker == "]":
+            level -= 1
+            if level == 0:
+                found = True
+                break
+
+        prevPos = state.pos
+        state.md.inline.skipToken(state)
+        if marker == "[":
+            if prevPos == state.pos - 1:
+                # increase level if we find text `[`,
+                # which is not a part of any token
+                level += 1
+            elif disableNested:
+                state.pos = oldPos
+                return -1
+    if found:
+        labelEnd = state.pos
+
+    # restore old state
+    state.pos = oldPos
+
+    return labelEnd

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/helpers/parse_link_title.py

@@ -0,0 +1,75 @@
+"""Parse link title"""
+
+from ..common.utils import charCodeAt, unescapeAll
+
+
+class _State:
+    __slots__ = ("can_continue", "marker", "ok", "pos", "str")
+
+    def __init__(self) -> None:
+        self.ok = False
+        """if `true`, this is a valid link title"""
+        self.can_continue = False
+        """if `true`, this link can be continued on the next line"""
+        self.pos = 0
+        """if `ok`, it's the position of the first character after the closing marker"""
+        self.str = ""
+        """if `ok`, it's the unescaped title"""
+        self.marker = 0
+        """expected closing marker character code"""
+
+    def __str__(self) -> str:
+        return self.str
+
+
+def parseLinkTitle(
+    string: str, start: int, maximum: int, prev_state: _State | None = None
+) -> _State:
+    """Parse link title within `str` in [start, max] range,
+    or continue previous parsing if `prev_state` is defined (equal to result of last execution).
+    """
+    pos = start
+    state = _State()
+
+    if prev_state is not None:
+        # this is a continuation of a previous parseLinkTitle call on the next line,
+        # used in reference links only
+        state.str = prev_state.str
+        state.marker = prev_state.marker
+    else:
+        if pos >= maximum:
+            return state
+
+        marker = charCodeAt(string, pos)
+
+        # /* " */  /* ' */  /* ( */
+        if marker != 0x22 and marker != 0x27 and marker != 0x28:
+            return state
+
+        start += 1
+        pos += 1
+
+        # if opening marker is "(", switch it to closing marker ")"
+        if marker == 0x28:
+            marker = 0x29
+
+        state.marker = marker
+
+    while pos < maximum:
+        code = charCodeAt(string, pos)
+        if code == state.marker:
+            state.pos = pos + 1
+            state.str += unescapeAll(string[start:pos])
+            state.ok = True
+            return state
+        elif code == 0x28 and state.marker == 0x29:  # /* ( */  /* ) */
+            return state
+        elif code == 0x5C and pos + 1 < maximum:  # /* \ */
+            pos += 1
+
+        pos += 1
+
+    # no closing marker found, but this link title may continue on the next line (for references)
+    state.can_continue = True
+    state.str += unescapeAll(string[start:pos])
+    return state

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/presets/__init__.py

@@ -0,0 +1,48 @@
+__all__ = ("commonmark", "default", "gfm_like", "gfm_like2", "js_default", "zero")
+
+from ..utils import PresetType
+from . import commonmark, default, zero
+
+js_default = default
+
+
+class gfm_like:  # noqa: N801
+    """GitHub Flavoured Markdown (GFM) like.
+
+    This adds the linkify, table and strikethrough components to CommmonMark.
+
+    Note, it lacks task-list items and raw HTML filtering,
+    to meet the the full GFM specification
+    (see https://github.github.com/gfm/#autolinks-extension-).
+    """
+
+    @staticmethod
+    def make() -> PresetType:
+        config = commonmark.make()
+        config["components"]["core"]["rules"].append("linkify")
+        config["components"]["block"]["rules"].append("table")
+        config["components"]["inline"]["rules"].extend(["strikethrough", "linkify"])
+        config["components"]["inline"]["rules2"].append("strikethrough")
+        config["options"]["linkify"] = True
+        config["options"]["html"] = True
+        return config
+
+
+class gfm_like2:  # noqa: N801
+    """GitHub Flavoured Markdown (GFM) like, extended.
+
+    Builds on ``gfm-like`` and additionally enables:
+
+    - Task lists (``- [x] done``)
+    - Alerts (``> [!NOTE]``)
+    - Single-tilde strikethrough (``~text~`` in addition to ``~~text~~``)
+    """
+
+    @staticmethod
+    def make() -> PresetType:
+        config = gfm_like.make()
+        config["options"]["tasklists"] = True
+        config["options"]["tasklists_editable"] = False
+        config["options"]["alerts"] = True
+        config["options"]["strikethrough_single_tilde"] = True
+        return config

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/presets/commonmark.py

@@ -0,0 +1,75 @@
+"""Commonmark default options.
+
+This differs to presets.default,
+primarily in that it allows HTML and does not enable components:
+
+- block: table
+- inline: strikethrough
+"""
+
+from ..utils import PresetType
+
+
+def make() -> PresetType:
+    return {
+        "options": {
+            "maxNesting": 20,  # Internal protection, recursion limit
+            "html": True,  # Enable HTML tags in source,
+            # this is just a shorthand for .enable(["html_inline", "html_block"])
+            # used by the linkify rule:
+            "linkify": False,  # autoconvert URL-like texts to links
+            # used by the replacements and smartquotes rules
+            # Enable some language-neutral replacements + quotes beautification
+            "typographer": False,
+            # used by the smartquotes rule:
+            # Double + single quotes replacement pairs, when typographer enabled,
+            # and smartquotes on. Could be either a String or an Array.
+            #
+            # For example, you can use '«»„“' for Russian, '„“‚‘' for German,
+            # and ['«\xA0', '\xA0»', '‹\xA0', '\xA0›'] for French (including nbsp).
+            "quotes": "\u201c\u201d\u2018\u2019",  # /* “”‘’ */
+            # Renderer specific; these options are used directly in the HTML renderer
+            "xhtmlOut": True,  # Use '/' to close single tags (<br />)
+            "breaks": False,  # Convert '\n' in paragraphs into <br>
+            "langPrefix": "language-",  # CSS language prefix for fenced blocks
+            # Highlighter function. Should return escaped HTML,
+            # or '' if the source string is not changed and should be escaped externally.
+            # If result starts with <pre... internal wrapper is skipped.
+            #
+            # function (/*str, lang, attrs*/) { return ''; }
+            #
+            "highlight": None,
+        },
+        "components": {
+            "core": {"rules": ["normalize", "block", "inline", "text_join"]},
+            "block": {
+                "rules": [
+                    "blockquote",
+                    "code",
+                    "fence",
+                    "heading",
+                    "hr",
+                    "html_block",
+                    "lheading",
+                    "list",
+                    "reference",
+                    "paragraph",
+                ]
+            },
+            "inline": {
+                "rules": [
+                    "autolink",
+                    "backticks",
+                    "emphasis",
+                    "entity",
+                    "escape",
+                    "html_inline",
+                    "image",
+                    "link",
+                    "newline",
+                    "text",
+                ],
+                "rules2": ["balance_pairs", "emphasis", "fragments_join"],
+            },
+        },
+    }

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/presets/default.py

@@ -0,0 +1,36 @@
+"""markdown-it default options."""
+
+from ..utils import PresetType
+
+
+def make() -> PresetType:
+    return {
+        "options": {
+            "maxNesting": 100,  # Internal protection, recursion limit
+            "html": False,  # Enable HTML tags in source
+            # this is just a shorthand for .disable(["html_inline", "html_block"])
+            # used by the linkify rule:
+            "linkify": False,  # autoconvert URL-like texts to links
+            # used by the replacements and smartquotes rules:
+            # Enable some language-neutral replacements + quotes beautification
+            "typographer": False,
+            # used by the smartquotes rule:
+            # Double + single quotes replacement pairs, when typographer enabled,
+            # and smartquotes on. Could be either a String or an Array.
+            # For example, you can use '«»„“' for Russian, '„“‚‘' for German,
+            # and ['«\xA0', '\xA0»', '‹\xA0', '\xA0›'] for French (including nbsp).
+            "quotes": "\u201c\u201d\u2018\u2019",  # /* “”‘’ */
+            # Renderer specific; these options are used directly in the HTML renderer
+            "xhtmlOut": False,  # Use '/' to close single tags (<br />)
+            "breaks": False,  # Convert '\n' in paragraphs into <br>
+            "langPrefix": "language-",  # CSS language prefix for fenced blocks
+            # Highlighter function. Should return escaped HTML,
+            # or '' if the source string is not changed and should be escaped externally.
+            # If result starts with <pre... internal wrapper is skipped.
+            #
+            # function (/*str, lang, attrs*/) { return ''; }
+            #
+            "highlight": None,
+        },
+        "components": {"core": {}, "block": {}, "inline": {}},
+    }

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/presets/zero.py

@@ -0,0 +1,44 @@
+"""
+"Zero" preset, with nothing enabled. Useful for manual configuring of simple
+modes. For example, to parse bold/italic only.
+"""
+
+from ..utils import PresetType
+
+
+def make() -> PresetType:
+    return {
+        "options": {
+            "maxNesting": 20,  # Internal protection, recursion limit
+            "html": False,  # Enable HTML tags in source
+            # this is just a shorthand for .disable(["html_inline", "html_block"])
+            # used by the linkify rule:
+            "linkify": False,  # autoconvert URL-like texts to links
+            # used by the replacements and smartquotes rules:
+            # Enable some language-neutral replacements + quotes beautification
+            "typographer": False,
+            # used by the smartquotes rule:
+            # Double + single quotes replacement pairs, when typographer enabled,
+            # and smartquotes on. Could be either a String or an Array.
+            # For example, you can use '«»„“' for Russian, '„“‚‘' for German,
+            # and ['«\xA0', '\xA0»', '‹\xA0', '\xA0›'] for French (including nbsp).
+            "quotes": "\u201c\u201d\u2018\u2019",  # /* “”‘’ */
+            # Renderer specific; these options are used directly in the HTML renderer
+            "xhtmlOut": False,  # Use '/' to close single tags (<br />)
+            "breaks": False,  # Convert '\n' in paragraphs into <br>
+            "langPrefix": "language-",  # CSS language prefix for fenced blocks
+            # Highlighter function. Should return escaped HTML,
+            # or '' if the source string is not changed and should be escaped externally.
+            # If result starts with <pre... internal wrapper is skipped.
+            # function (/*str, lang, attrs*/) { return ''; }
+            "highlight": None,
+        },
+        "components": {
+            "core": {"rules": ["normalize", "block", "inline", "text_join"]},
+            "block": {"rules": ["paragraph"]},
+            "inline": {
+                "rules": ["text"],
+                "rules2": ["balance_pairs", "fragments_join"],
+            },
+        },
+    }

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/rules_block/blockquote.py

@@ -0,0 +1,368 @@
+# Block quotes
+from __future__ import annotations
+
+import logging
+
+from ..common.utils import isStrSpace
+from .state_block import StateBlock
+
+LOGGER = logging.getLogger(__name__)
+
+
+def blockquote(state: StateBlock, startLine: int, endLine: int, silent: bool) -> bool:
+    LOGGER.debug(
+        "entering blockquote: %s, %s, %s, %s", state, startLine, endLine, silent
+    )
+
+    oldLineMax = state.lineMax
+    pos = state.bMarks[startLine] + state.tShift[startLine]
+    max = state.eMarks[startLine]
+
+    if state.is_code_block(startLine):
+        return False
+
+    # check the block quote marker
+    try:
+        if state.src[pos] != ">":
+            return False
+    except IndexError:
+        return False
+    pos += 1
+
+    # we know that it's going to be a valid blockquote,
+    # so no point trying to find the end of it in silent mode
+    if silent:
+        return True
+
+    # set offset past spaces and ">"
+    initial = offset = state.sCount[startLine] + 1
+
+    try:
+        second_char: str | None = state.src[pos]
+    except IndexError:
+        second_char = None
+
+    # skip one optional space after '>'
+    if second_char == " ":
+        # ' >   test '
+        #     ^ -- position start of line here:
+        pos += 1
+        initial += 1
+        offset += 1
+        adjustTab = False
+        spaceAfterMarker = True
+    elif second_char == "\t":
+        spaceAfterMarker = True
+
+        if (state.bsCount[startLine] + offset) % 4 == 3:
+            # '  >\t  test '
+            #       ^ -- position start of line here (tab has width==1)
+            pos += 1
+            initial += 1
+            offset += 1
+            adjustTab = False
+        else:
+            # ' >\t  test '
+            #    ^ -- position start of line here + shift bsCount slightly
+            #         to make extra space appear
+            adjustTab = True
+
+    else:
+        spaceAfterMarker = False
+
+    oldBMarks = [state.bMarks[startLine]]
+    state.bMarks[startLine] = pos
+
+    while pos < max:
+        ch = state.src[pos]
+
+        if isStrSpace(ch):
+            if ch == "\t":
+                offset += (
+                    4
+                    - (offset + state.bsCount[startLine] + (1 if adjustTab else 0)) % 4
+                )
+            else:
+                offset += 1
+
+        else:
+            break
+
+        pos += 1
+
+    oldBSCount = [state.bsCount[startLine]]
+    state.bsCount[startLine] = (
+        state.sCount[startLine] + 1 + (1 if spaceAfterMarker else 0)
+    )
+
+    lastLineEmpty = pos >= max
+
+    oldSCount = [state.sCount[startLine]]
+    state.sCount[startLine] = offset - initial
+
+    oldTShift = [state.tShift[startLine]]
+    state.tShift[startLine] = pos - state.bMarks[startLine]
+
+    terminatorRules = state.md.block.ruler.getRules("blockquote")
+
+    oldParentType = state.parentType
+    state.parentType = "blockquote"
+
+    # Search the end of the block
+    #
+    # Block ends with either:
+    #  1. an empty line outside:
+    #     ```
+    #     > test
+    #
+    #     ```
+    #  2. an empty line inside:
+    #     ```
+    #     >
+    #     test
+    #     ```
+    #  3. another tag:
+    #     ```
+    #     > test
+    #      - - -
+    #     ```
+
+    # for (nextLine = startLine + 1; nextLine < endLine; nextLine++) {
+    nextLine = startLine + 1
+    while nextLine < endLine:
+        # check if it's outdented, i.e. it's inside list item and indented
+        # less than said list item:
+        #
+        # ```
+        # 1. anything
+        #    > current blockquote
+        # 2. checking this line
+        # ```
+        isOutdented = state.sCount[nextLine] < state.blkIndent
+
+        pos = state.bMarks[nextLine] + state.tShift[nextLine]
+        max = state.eMarks[nextLine]
+
+        if pos >= max:
+            # Case 1: line is not inside the blockquote, and this line is empty.
+            break
+
+        evaluatesTrue = state.src[pos] == ">" and not isOutdented
+        pos += 1
+        if evaluatesTrue:
+            # This line is inside the blockquote.
+
+            # set offset past spaces and ">"
+            initial = offset = state.sCount[nextLine] + 1
+
+            try:
+                next_char: str | None = state.src[pos]
+            except IndexError:
+                next_char = None
+
+            # skip one optional space after '>'
+            if next_char == " ":
+                # ' >   test '
+                #     ^ -- position start of line here:
+                pos += 1
+                initial += 1
+                offset += 1
+                adjustTab = False
+                spaceAfterMarker = True
+            elif next_char == "\t":
+                spaceAfterMarker = True
+
+                if (state.bsCount[nextLine] + offset) % 4 == 3:
+                    # '  >\t  test '
+                    #       ^ -- position start of line here (tab has width==1)
+                    pos += 1
+                    initial += 1
+                    offset += 1
+                    adjustTab = False
+                else:
+                    # ' >\t  test '
+                    #    ^ -- position start of line here + shift bsCount slightly
+                    #         to make extra space appear
+                    adjustTab = True
+
+            else:
+                spaceAfterMarker = False
+
+            oldBMarks.append(state.bMarks[nextLine])
+            state.bMarks[nextLine] = pos
+
+            while pos < max:
+                ch = state.src[pos]
+
+                if isStrSpace(ch):
+                    if ch == "\t":
+                        offset += (
+                            4
+                            - (
+                                offset
+                                + state.bsCount[nextLine]
+                                + (1 if adjustTab else 0)
+                            )
+                            % 4
+                        )
+                    else:
+                        offset += 1
+                else:
+                    break
+
+                pos += 1
+
+            lastLineEmpty = pos >= max
+
+            oldBSCount.append(state.bsCount[nextLine])
+            state.bsCount[nextLine] = (
+                state.sCount[nextLine] + 1 + (1 if spaceAfterMarker else 0)
+            )
+
+            oldSCount.append(state.sCount[nextLine])
+            state.sCount[nextLine] = offset - initial
+
+            oldTShift.append(state.tShift[nextLine])
+            state.tShift[nextLine] = pos - state.bMarks[nextLine]
+
+            nextLine += 1
+            continue
+
+        # Case 2: line is not inside the blockquote, and the last line was empty.
+        if lastLineEmpty:
+            break
+
+        # Case 3: another tag found.
+        terminate = False
+
+        for terminatorRule in terminatorRules:
+            if terminatorRule(state, nextLine, endLine, True):
+                terminate = True
+                break
+
+        if terminate:
+            # Quirk to enforce "hard termination mode" for paragraphs;
+            # normally if you call `tokenize(state, startLine, nextLine)`,
+            # paragraphs will look below nextLine for paragraph continuation,
+            # but if blockquote is terminated by another tag, they shouldn't
+            state.lineMax = nextLine
+
+            if state.blkIndent != 0:
+                # state.blkIndent was non-zero, we now set it to zero,
+                # so we need to re-calculate all offsets to appear as
+                # if indent wasn't changed
+                oldBMarks.append(state.bMarks[nextLine])
+                oldBSCount.append(state.bsCount[nextLine])
+                oldTShift.append(state.tShift[nextLine])
+                oldSCount.append(state.sCount[nextLine])
+                state.sCount[nextLine] -= state.blkIndent
+
+            break
+
+        oldBMarks.append(state.bMarks[nextLine])
+        oldBSCount.append(state.bsCount[nextLine])
+        oldTShift.append(state.tShift[nextLine])
+        oldSCount.append(state.sCount[nextLine])
+
+        # A negative indentation means that this is a paragraph continuation
+        #
+        state.sCount[nextLine] = -1
+
+        nextLine += 1
+
+    oldIndent = state.blkIndent
+    state.blkIndent = 0
+
+    # Detect GitHub-style alert marker on the first content line.
+    # Note: `startLine` here refers to the first content line of the
+    # blockquote, after the `>` prefix has already been stripped by the
+    # blockquote parser above (bMarks/tShift adjusted to skip `> `).
+    alert_kind = None
+    if state.md.options.get("alerts", False) and nextLine > startLine:
+        alert_kind = _detect_alert(state, startLine)
+
+    lines = [startLine, 0]
+
+    if alert_kind is not None:
+        # Emit alert tokens instead of blockquote tokens
+        alert_lower = alert_kind.lower()
+        token = state.push("alert_open", "div", 1)
+        token.markup = ">"
+        token.attrSet("class", f"markdown-alert markdown-alert-{alert_lower}")
+        token.map = lines
+        token.info = alert_kind
+        token.meta = {"kind": alert_kind}
+
+        # Emit a title paragraph: <p class="markdown-alert-title">Kind</p>
+        token = state.push("alert_title_open", "p", 1)
+        token.attrSet("class", "markdown-alert-title")
+        title_token = state.push("inline", "", 0)
+        title_token.content = alert_kind.capitalize()
+        title_token.children = []
+        token = state.push("alert_title_close", "p", -1)
+
+        # Skip the marker line (startLine) and tokenize from startLine + 1.
+        contentStart = startLine + 1
+        if contentStart < nextLine:
+            # tokenize() updates state.line to nextLine as part of its
+            # contract, consistent with the blockquote code path below.
+            state.md.block.tokenize(state, contentStart, nextLine)
+        else:
+            state.line = nextLine
+
+        token = state.push("alert_close", "div", -1)
+        token.markup = ">"
+    else:
+        token = state.push("blockquote_open", "blockquote", 1)
+        token.markup = ">"
+        token.map = lines
+
+        state.md.block.tokenize(state, startLine, nextLine)
+
+        token = state.push("blockquote_close", "blockquote", -1)
+        token.markup = ">"
+
+    state.lineMax = oldLineMax
+    state.parentType = oldParentType
+    # Update the opening token map for both alert and blockquote containers.
+    lines[1] = state.line
+
+    # Restore original tShift; this might not be necessary since the parser
+    # has already been here, but just to make sure we can do that.
+    for i, item in enumerate(oldTShift):
+        state.bMarks[i + startLine] = oldBMarks[i]
+        state.tShift[i + startLine] = item
+        state.sCount[i + startLine] = oldSCount[i]
+        state.bsCount[i + startLine] = oldBSCount[i]
+
+    state.blkIndent = oldIndent
+
+    return True
+
+
+_ALERT_TYPES = {"NOTE", "TIP", "IMPORTANT", "WARNING", "CAUTION"}
+
+
+def _detect_alert(state: StateBlock, startLine: int) -> str | None:
+    """Detect ``[!TYPE]`` on *startLine* (after ``>`` prefix has been stripped).
+
+    Returns the alert type string (e.g. ``"NOTE"``) or ``None``.
+    """
+    pos = state.bMarks[startLine] + state.tShift[startLine]
+    maximum = state.eMarks[startLine]
+    src = state.src
+
+    # Trim trailing whitespace
+    while maximum > pos and src[maximum - 1] in (" ", "\t"):
+        maximum -= 1
+
+    if maximum - pos < 4:
+        return None
+    if src[pos] != "[" or src[pos + 1] != "!":
+        return None
+    if src[maximum - 1] != "]":
+        return None
+    type_str = src[pos + 2 : maximum - 1].upper()
+    if type_str not in _ALERT_TYPES:
+        return None
+    return type_str

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/rules_block/heading.py

@@ -0,0 +1,69 @@
+"""Atex heading (#, ##, ...)"""
+
+from __future__ import annotations
+
+import logging
+
+from ..common.utils import isStrSpace
+from .state_block import StateBlock
+
+LOGGER = logging.getLogger(__name__)
+
+
+def heading(state: StateBlock, startLine: int, endLine: int, silent: bool) -> bool:
+    LOGGER.debug("entering heading: %s, %s, %s, %s", state, startLine, endLine, silent)
+
+    pos = state.bMarks[startLine] + state.tShift[startLine]
+    maximum = state.eMarks[startLine]
+
+    if state.is_code_block(startLine):
+        return False
+
+    ch: str | None = state.src[pos]
+
+    if ch != "#" or pos >= maximum:
+        return False
+
+    # count heading level
+    level = 1
+    pos += 1
+    try:
+        ch = state.src[pos]
+    except IndexError:
+        ch = None
+    while ch == "#" and pos < maximum and level <= 6:
+        level += 1
+        pos += 1
+        try:
+            ch = state.src[pos]
+        except IndexError:
+            ch = None
+
+    if level > 6 or (pos < maximum and not isStrSpace(ch)):
+        return False
+
+    if silent:
+        return True
+
+    # Let's cut tails like '    ###  ' from the end of string
+
+    maximum = state.skipSpacesBack(maximum, pos)
+    tmp = state.skipCharsStrBack(maximum, "#", pos)
+    if tmp > pos and isStrSpace(state.src[tmp - 1]):
+        maximum = tmp
+
+    state.line = startLine + 1
+
+    token = state.push("heading_open", "h" + str(level), 1)
+    token.markup = "########"[:level]
+    token.map = [startLine, state.line]
+
+    token = state.push("inline", "", 0)
+    token.content = state.src[pos:maximum].strip()
+    token.map = [startLine, state.line]
+    token.children = []
+
+    token = state.push("heading_close", "h" + str(level), -1)
+    token.markup = "########"[:level]
+
+    return True

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/rules_block/lheading.py

@@ -0,0 +1,86 @@
+# lheading (---, ==)
+import logging
+
+from .state_block import StateBlock
+
+LOGGER = logging.getLogger(__name__)
+
+
+def lheading(state: StateBlock, startLine: int, endLine: int, silent: bool) -> bool:
+    LOGGER.debug("entering lheading: %s, %s, %s, %s", state, startLine, endLine, silent)
+
+    level = None
+    nextLine = startLine + 1
+    ruler = state.md.block.ruler
+    terminatorRules = ruler.getRules("paragraph")
+
+    if state.is_code_block(startLine):
+        return False
+
+    oldParentType = state.parentType
+    state.parentType = "paragraph"  # use paragraph to match terminatorRules
+
+    # jump line-by-line until empty one or EOF
+    while nextLine < endLine and not state.isEmpty(nextLine):
+        # this would be a code block normally, but after paragraph
+        # it's considered a lazy continuation regardless of what's there
+        if state.sCount[nextLine] - state.blkIndent > 3:
+            nextLine += 1
+            continue
+
+        # Check for underline in setext header
+        if state.sCount[nextLine] >= state.blkIndent:
+            pos = state.bMarks[nextLine] + state.tShift[nextLine]
+            maximum = state.eMarks[nextLine]
+
+            if pos < maximum:
+                marker = state.src[pos]
+
+                if marker in ("-", "="):
+                    pos = state.skipCharsStr(pos, marker)
+                    pos = state.skipSpaces(pos)
+
+                    # /* = */
+                    if pos >= maximum:
+                        level = 1 if marker == "=" else 2
+                        break
+
+        # quirk for blockquotes, this line should already be checked by that rule
+        if state.sCount[nextLine] < 0:
+            nextLine += 1
+            continue
+
+        # Some tags can terminate paragraph without empty line.
+        terminate = False
+        for terminatorRule in terminatorRules:
+            if terminatorRule(state, nextLine, endLine, True):
+                terminate = True
+                break
+        if terminate:
+            break
+
+        nextLine += 1
+
+    if not level:
+        # Didn't find valid underline
+        return False
+
+    content = state.getLines(startLine, nextLine, state.blkIndent, False).strip()
+
+    state.line = nextLine + 1
+
+    token = state.push("heading_open", "h" + str(level), 1)
+    token.markup = marker
+    token.map = [startLine, state.line]
+
+    token = state.push("inline", "", 0)
+    token.content = content
+    token.map = [startLine, state.line - 1]
+    token.children = []
+
+    token = state.push("heading_close", "h" + str(level), -1)
+    token.markup = marker
+
+    state.parentType = oldParentType
+
+    return True

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/markdown_it/rules_block/reference.py

@@ -0,0 +1,235 @@
+import logging
+
+from ..common.utils import charCodeAt, isSpace, normalizeReference
+from .state_block import StateBlock
+
+LOGGER = logging.getLogger(__name__)
+
+
+def reference(state: StateBlock, startLine: int, _endLine: int, silent: bool) -> bool:
+    LOGGER.debug(
+        "entering reference: %s, %s, %s, %s", state, startLine, _endLine, silent
+    )
+
+    pos = state.bMarks[startLine] + state.tShift[startLine]
+    maximum = state.eMarks[startLine]
+    nextLine = startLine + 1
+
+    if state.is_code_block(startLine):
+        return False
+
+    if state.src[pos] != "[":
+        return False
+
+    string = state.src[pos : maximum + 1]
+
+    # string = state.getLines(startLine, nextLine, state.blkIndent, False).strip()
+    maximum = len(string)
+
+    labelEnd = None
+    pos = 1
+    while pos < maximum:
+        ch = charCodeAt(string, pos)
+        if ch == 0x5B:  # /* [ */
+            return False
+        elif ch == 0x5D:  # /* ] */
+            labelEnd = pos
+            break
+        elif ch == 0x0A:  # /* \n */
+            if (lineContent := getNextLine(state, nextLine)) is not None:
+                string += lineContent
+                maximum = len(string)
+                nextLine += 1
+        elif ch == 0x5C:  # /* \ */
+            pos += 1
+            if (
+                pos < maximum
+                and charCodeAt(string, pos) == 0x0A
+                and (lineContent := getNextLine(state, nextLine)) is not None
+            ):
+                string += lineContent
+                maximum = len(string)
+                nextLine += 1
+        pos += 1
+
+    if (
+        labelEnd is None or labelEnd < 0 or charCodeAt(string, labelEnd + 1) != 0x3A
+    ):  # /* : */
+        return False
+
+    # [label]:   destination   'title'
+    #         ^^^ skip optional whitespace here
+    pos = labelEnd + 2
+    while pos < maximum:
+        ch = charCodeAt(string, pos)
+        if ch == 0x0A:
+            if (lineContent := getNextLine(state, nextLine)) is not None:
+                string += lineContent
+                maximum = len(string)
+                nextLine += 1
+        elif isSpace(ch):
+            pass
+        else:
+            break
+        pos += 1
+
+    # [label]:   destination   'title'
+    #            ^^^^^^^^^^^ parse this
+    destRes = state.md.helpers.parseLinkDestination(string, pos, maximum)
+    if not destRes.ok:
+        return False
+
+    href = state.md.normalizeLink(destRes.str)
+    if not state.md.validateLink(href):
+        return False
+
+    pos = destRes.pos
+
+    # save cursor state, we could require to rollback later
+    destEndPos = pos
+    destEndLineNo = nextLine
+
+    # [label]:   destination   'title'
+    #                       ^^^ skipping those spaces
+    start = pos
+    while pos < maximum:
+        ch = charCodeAt(string, pos)
+        if ch == 0x0A:
+            if (lineContent := getNextLine(state, nextLine)) is not None:
+                string += lineContent
+                maximum = len(string)
+                nextLine += 1
+        elif isSpace(ch):
+            pass
+        else:
+            break
+        pos += 1
+
+    # [label]:   destination   'title'
+    #                          ^^^^^^^ parse this
+    titleRes = state.md.helpers.parseLinkTitle(string, pos, maximum, None)
+    while titleRes.can_continue:
+        if (lineContent := getNextLine(state, nextLine)) is None:
+            break
+        string += lineContent
+        pos = maximum
+        maximum = len(string)
+        nextLine += 1
+        titleRes = state.md.helpers.parseLinkTitle(string, pos, maximum, titleRes)
+
+    if pos < maximum and start != pos and titleRes.ok:
+        title = titleRes.str
+        pos = titleRes.pos
+    else:
+        title = ""
+        pos = destEndPos
+        nextLine = destEndLineNo
+
+    # skip trailing spaces until the rest of the line
+    while pos < maximum:
+        ch = charCodeAt(string, pos)
+        if not isSpace(ch):
+            break
+        pos += 1
+
+    if pos < maximum and charCodeAt(string, pos) != 0x0A and title:
+        # garbage at the end of the line after title,
+        # but it could still be a valid reference if we roll back
+        title = ""
+        pos = destEndPos
+        nextLine = destEndLineNo
+        while pos < maximum:
+            ch = charCodeAt(string, pos)
+            if not isSpace(ch):
+                break
+            pos += 1
+
+    if pos < maximum and charCodeAt(string, pos) != 0x0A:
+        # garbage at the end of the line
+        return False
+
+    label = normalizeReference(string[1:labelEnd])
+    if not label:
+        # CommonMark 0.20 disallows empty labels
+        return False
+
+    # Reference can not terminate anything. This check is for safety only.
+    if silent:
+        return True
+
+    if "references" not in state.env:
+        state.env["references"] = {}
+
+    state.line = nextLine
+
+    # note, this is not part of markdown-it JS, but is useful for renderers
+    if state.md.options.get("inline_definitions", False):
+        token = state.push("definition", "", 0)
+        token.meta = {
+            "id": label,
+            "title": title,
+            "url": href,
+            "label": string[1:labelEnd],
+        }
+        token.map = [startLine, state.line]
+
+    if label not in state.env["references"]:
+        state.env["references"][label] = {
+            "title": title,
+            "href": href,
+            "map": [startLine, state.line],
+        }
+    else:
+        state.env.setdefault("duplicate_refs", []).append(
+            {
+                "title": title,
+                "href": href,
+                "label": label,
+                "map": [startLine, state.line],
+            }
+        )
+
+    return True
+
+
+def getNextLine(state: StateBlock, nextLine: int) -> None | str:
+    endLine = state.lineMax
+
+    if nextLine >= endLine or state.isEmpty(nextLine):
+        # empty line or end of input
+        return None
+
+    isContinuation = False
+
+    # this would be a code block normally, but after paragraph
+    # it's considered a lazy continuation regardless of what's there
+    if state.is_code_block(nextLine):
+        isContinuation = True
+
+    # quirk for blockquotes, this line should already be checked by that rule
+    if state.sCount[nextLine] < 0:
+        isContinuation = True
+
+    if not isContinuation:
+        terminatorRules = state.md.block.ruler.getRules("reference")
+        oldParentType = state.parentType
+        state.parentType = "reference"
+
+        # Some tags can terminate paragraph without empty line.
+        terminate = False
+        for terminatorRule in terminatorRules:
+            if terminatorRule(state, nextLine, endLine, True):
+                terminate = True
+                break
+
+        state.parentType = oldParentType
+
+        if terminate:
+            # terminated by another block
+            return None
+
+    pos = state.bMarks[nextLine] + state.tShift[nextLine]
+    maximum = state.eMarks[nextLine]
+
+    # max + 1 explicitly includes the newline
+    return state.src[pos : maximum + 1]

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/transformers/models/altclip/__init__.py

@@ -0,0 +1,28 @@
+# Copyright 2020 The HuggingFace Team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+from typing import TYPE_CHECKING
+
+from ...utils import _LazyModule
+from ...utils.import_utils import define_import_structure
+
+
+if TYPE_CHECKING:
+    from .configuration_altclip import *
+    from .modeling_altclip import *
+    from .processing_altclip import *
+else:
+    import sys
+
+    _file = globals()["__file__"]
+    sys.modules[__name__] = _LazyModule(__name__, _file, define_import_structure(_file), module_spec=__spec__)

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/transformers/models/altclip/configuration_altclip.py

@@ -0,0 +1,246 @@
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+#           This file was automatically generated from src/transformers/models/altclip/modular_altclip.py.
+#               Do NOT edit this file manually as any edits will be overwritten by the generation of
+#             the file from the modular. If any change should be done, please apply the change to the
+#                          modular_altclip.py file directly. One of our CI enforces this.
+#                🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
+# Copyright 2022 The BAAI Teams Authors and The HuggingFace Inc. team. All rights reserved.
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+from huggingface_hub.dataclasses import strict
+
+from ...configuration_utils import PreTrainedConfig
+from ...utils import auto_docstring, logging
+
+
+logger = logging.get_logger(__name__)
+
+
+@auto_docstring(checkpoint="BAAI/AltCLIP")
+@strict
+class AltCLIPTextConfig(PreTrainedConfig):
+    r"""
+    project_dim (`int`, *optional*, defaults to 768):
+        The dimensions of the teacher model before the mapping layer.
+
+    Examples:
+
+    ```python
+    >>> from transformers import AltCLIPTextModel, AltCLIPTextConfig
+
+    >>> # Initializing a AltCLIPTextConfig with BAAI/AltCLIP style configuration
+    >>> configuration = AltCLIPTextConfig()
+
+    >>> # Initializing a AltCLIPTextModel (with random weights) from the BAAI/AltCLIP style configuration
+    >>> model = AltCLIPTextModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "altclip_text_model"
+    base_config_key = "text_config"
+
+    vocab_size: int = 250002
+    hidden_size: int = 1024
+    intermediate_size: int = 4096
+    num_hidden_layers: int = 24
+    num_attention_heads: int = 16
+    max_position_embeddings: int = 514
+    hidden_act: str = "gelu"
+    layer_norm_eps: float | None = 1e-5
+    initializer_range: float = 0.02
+    initializer_factor: float = 0.02
+    pad_token_id: int | None = 1
+    bos_token_id: int | None = 0
+    eos_token_id: int | None = 2
+    hidden_dropout_prob: int | float = 0.1
+    attention_probs_dropout_prob: int | float = 0
+    type_vocab_size: int = 1
+    project_dim: int = 768
+
+    def validate_architecture(self):
+        """Part of `@strict`-powered validation. Validates the architecture of the config."""
+        if self.hidden_size % self.num_attention_heads != 0:
+            raise ValueError(
+                f"The hidden size ({self.hidden_size}) is not a multiple of the number of attention "
+                f"heads ({self.num_attention_heads})."
+            )
+
+
+@auto_docstring(checkpoint="BAAI/AltCLIP")
+@strict
+class AltCLIPVisionConfig(PreTrainedConfig):
+    r"""
+    Example:
+
+    ```python
+    >>> from transformers import AltCLIPVisionConfig, AltCLIPVisionModel
+
+    >>> # Initializing a AltCLIPVisionConfig with BAAI/AltCLIP style configuration
+    >>> configuration = AltCLIPVisionConfig()
+
+    >>> # Initializing a AltCLIPVisionModel (with random weights) from the BAAI/AltCLIP style configuration
+    >>> model = AltCLIPVisionModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+
+    model_type = "altclip_vision_model"
+    base_config_key = "vision_config"
+
+    hidden_size: int = 768
+    intermediate_size: int = 3072
+    projection_dim: int = 512
+    num_hidden_layers: int = 12
+    num_attention_heads: int = 12
+    num_channels: int = 3
+    image_size: int | list[int] | tuple[int, int] | None = 224
+    patch_size: int | list[int] | tuple[int, int] | None = 32
+    hidden_act: str = "quick_gelu"
+    layer_norm_eps: float = 1e-5
+    attention_dropout: int | float | None = 0.0
+    initializer_range: float = 0.02
+    initializer_factor: float = 1.0
+
+    def validate_architecture(self):
+        """Part of `@strict`-powered validation. Validates the architecture of the config."""
+        if self.hidden_size % self.num_attention_heads != 0:
+            raise ValueError(
+                f"The hidden size ({self.hidden_size}) is not a multiple of the number of attention "
+                f"heads ({self.num_attention_heads})."
+            )
+
+
+@auto_docstring(checkpoint="BAAI/AltCLIP")
+@strict
+class AltCLIPConfig(PreTrainedConfig):
+    r"""
+    Example:
+
+    ```python
+    >>> from transformers import AltCLIPConfig, AltCLIPModel
+
+    >>> # Initializing a AltCLIPConfig with BAAI/AltCLIP style configuration
+    >>> configuration = AltCLIPConfig()
+
+    >>> # Initializing a AltCLIPModel (with random weights) from the BAAI/AltCLIP style configuration
+    >>> model = AltCLIPModel(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+
+    >>> # We can also initialize a AltCLIPConfig from a AltCLIPTextConfig and a AltCLIPVisionConfig
+
+    >>> # Initializing a AltCLIPText and AltCLIPVision configuration
+    >>> config_text = AltCLIPTextConfig()
+    >>> config_vision = AltCLIPVisionConfig()
+
+    >>> config = AltCLIPConfig(text_config=config_text, vision_config=config_vision)
+    ```"""
+
+    model_type = "altclip"
+    sub_configs = {"text_config": AltCLIPTextConfig, "vision_config": AltCLIPVisionConfig}
+
+    text_config: dict | AltCLIPTextConfig | None = None
+    vision_config: dict | AltCLIPVisionConfig | None = None
+
+    projection_dim: int = 768
+    logit_scale_init_value: float | int | None = 2.6592
+    initializer_factor: float | None = 1.0
+
+    def __post_init__(self, **kwargs):
+        if self.text_config is None:
+            text_config = {}
+            logger.info("`text_config` is `None`. Initializing the `AltCLIPTextConfig` with default values.")
+        elif isinstance(self.text_config, AltCLIPTextConfig):
+            text_config = self.text_config.to_dict()
+        else:
+            text_config = self.text_config
+
+        if self.vision_config is None:
+            vision_config = {}
+            logger.info("`vision_config` is `None`. initializing the `AltCLIPVisionConfig` with default values.")
+        elif isinstance(self.vision_config, AltCLIPVisionConfig):
+            vision_config = self.vision_config.to_dict()
+        else:
+            vision_config = self.vision_config
+
+        # For backward compatibility check keyword args
+        # Instead of simply assigning `[text|vision]_config_dict` to `[text|vision]_config`, we use the values in
+        # `[text|vision]_config_dict` to update the values in `[text|vision]_config`. The values should be same in most
+        # cases, but we don't want to break anything regarding `_config_dict` that existed before commit `8827e1b2`.
+        text_config_dict = kwargs.pop("text_config_dict", None)
+        vision_config_dict = kwargs.pop("vision_config_dict", None)
+
+        if text_config_dict is not None:
+            # This is the complete result when using `text_config_dict`.
+            _text_config_dict = AltCLIPTextConfig(**text_config_dict).to_dict()
+
+            # Give a warning if the values exist in both `_text_config_dict` and `text_config` but being different.
+            for key, value in _text_config_dict.items():
+                if key in text_config and value != text_config[key] and key != "transformers_version":
+                    # If specified in `text_config_dict`
+                    if key in text_config_dict:
+                        message = (
+                            f"`{key}` is found in both `text_config_dict` and `text_config` but with different values. "
+                            f'The value `text_config_dict["{key}"]` will be used instead.'
+                        )
+                    # If inferred from default argument values (just to be super careful)
+                    else:
+                        message = (
+                            f"`text_config_dict` is provided which will be used to initialize `CLIPTextConfig`. The "
+                            f'value `text_config["{key}"]` will be overridden.'
+                        )
+                    logger.info(message)
+
+            # Update all values in `text_config` with the ones in `_text_config_dict`.
+            text_config.update(_text_config_dict)
+
+        if vision_config_dict is not None:
+            # This is the complete result when using `vision_config_dict`.
+            _vision_config_dict = AltCLIPVisionConfig(**vision_config_dict).to_dict()
+            # convert keys to string instead of integer
+            if "id2label" in _vision_config_dict:
+                _vision_config_dict["id2label"] = {
+                    str(key): value for key, value in _vision_config_dict["id2label"].items()
+                }
+
+            # Give a warning if the values exist in both `_vision_config_dict` and `vision_config` but being different.
+            for key, value in _vision_config_dict.items():
+                if key in vision_config and value != vision_config[key] and key != "transformers_version":
+                    # If specified in `vision_config_dict`
+                    if key in vision_config_dict:
+                        message = (
+                            f"`{key}` is found in both `vision_config_dict` and `vision_config` but with different "
+                            f'values. The value `vision_config_dict["{key}"]` will be used instead.'
+                        )
+                    # If inferred from default argument values (just to be super careful)
+                    else:
+                        message = (
+                            f"`vision_config_dict` is provided which will be used to initialize `CLIPVisionConfig`. "
+                            f'The value `vision_config["{key}"]` will be overridden.'
+                        )
+                    logger.info(message)
+
+            # Update all values in `vision_config` with the ones in `_vision_config_dict`.
+            vision_config.update(_vision_config_dict)
+
+        # Finally we can convert back our unified text/vision configs to `PretrainedConfig`
+        self.text_config = AltCLIPTextConfig(**text_config)
+        self.vision_config = AltCLIPVisionConfig(**vision_config)
+
+        super().__post_init__(**kwargs)
+
+
+__all__ = ["AltCLIPTextConfig", "AltCLIPVisionConfig", "AltCLIPConfig"]

LTA_openwebtext_dualt/mini_owt_logdirichlet/.venv_qwen35_uv/lib/python3.12/site-packages/transformers/models/altclip/processing_altclip.py

@@ -0,0 +1,28 @@
+# Copyright 2022 WenXiang ZhongzhiCheng LedellWu LiuGuang BoWenZhang The HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""
+Image/Text processor class for AltCLIP
+"""
+
+from ...processing_utils import ProcessorMixin
+from ...utils import auto_docstring
+
+
+@auto_docstring
+class AltCLIPProcessor(ProcessorMixin):
+    def __init__(self, image_processor=None, tokenizer=None):
+        super().__init__(image_processor, tokenizer)
+
+
+__all__ = ["AltCLIPProcessor"]

LTA_openwebtext_dualt/mini_owt_logdirichlet/runs/owt_t5_elftokenized_full_len1024_C1_to_1024_pow1_d768_l12_h12_gbs512_2x8gpu_50ep_lr4e3_ema0p9999_elfopt_not5_bottleneck128_unfixed_norm_stateprobadd_selfcond_ce_fast_20260610_020108/step_003000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa6d3de70bd52b21c01abd539fe39af93889121f3ac7bffb10d4cafa55882001
+size 927700322

LTA_openwebtext_dualt/mini_owt_logdirichlet/runs/owt_t5_elftokenized_full_len1024_C1_to_1024_pow1_d768_l12_h12_gbs512_2x8gpu_50ep_lr4e3_ema0p9999_elfopt_not5_bottleneck128_unfixed_norm_stateprobadd_selfcond_ce_fast_20260610_020108/step_179000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbc0a9693f5ee8aec606d5b7c2dfa61a4723f697272c48d06e3393ea93687f4d
+size 927700322

LTA_openwebtext_dualt/mini_owt_logdirichlet/runs/owt_t5_elftokenized_full_len1024_C1_to_1024_pow1_d768_l12_h12_gbs512_2x8gpu_50ep_lr4e3_ema0p9999_elfopt_not5_bottleneck128_unfixed_norm_stateprobadd_selfcond_ce_fast_20260610_020108/step_185000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60f25cc52272898fec4310310b99b1c39b8a05c025b64c1cc37107ca1d354afb
+size 927700322

LTA_openwebtext_dualt/mini_owt_logdirichlet/runs/owt_t5_elftokenized_full_len1024_C1_to_1024_pow1_d768_l12_h12_gbs512_2x8gpu_50ep_lr4e3_ema0p9999_elfopt_not5_bottleneck128_unfixed_norm_stateprobadd_selfcond_ce_fast_20260610_020108/step_192000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d90c3432ab622266946dd0cd272e89740e6f41214d5aab98a170b9172d88f8e2
+size 927700322

LTA_openwebtext_dualt/mini_owt_logdirichlet/runs/owt_t5_elftokenized_full_len1024_C1_to_1024_pow1_d768_l12_h12_gbs512_2x8gpu_50ep_lr4e3_ema0p9999_elfopt_not5_bottleneck128_unfixed_norm_stateprobadd_selfcond_ce_fast_20260610_020108/step_279000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64654fc51957676402ac4ab478eec65183ad55382beac9e45ff6df293bab67be
+size 927700322