{ "_name_or_path": "raomnb/SN388", "all_reduce_scores": { "0": "NOT_ALIVE", "1": "NOT_ALIVE", "10": "NOT_ALIVE", "100": "NON_PARTICIPATING", "101": "NOT_ALIVE", "102": "NON_PARTICIPATING", "103": "NON_PARTICIPATING", "104": "NON_PARTICIPATING", "105": "NON_PARTICIPATING", "106": "NOT_ALIVE", "107": "NON_PARTICIPATING", "108": "NON_PARTICIPATING", "109": "NOT_ALIVE", "11": "NOT_ALIVE", "110": "NOT_ALIVE", "111": "NON_PARTICIPATING", "112": "NON_PARTICIPATING", "113": "NOT_ALIVE", "114": "NON_PARTICIPATING", "115": "NON_PARTICIPATING", "116": "NOT_ALIVE", "117": "NOT_ALIVE", "118": "NOT_ALIVE", "119": "NON_PARTICIPATING", "12": "NOT_ALIVE", "120": "NON_PARTICIPATING", "121": "NON_PARTICIPATING", "122": "NOT_ALIVE", "123": "NOT_ALIVE", "124": "NON_PARTICIPATING", "125": "NOT_ALIVE", "126": "NOT_ALIVE", "127": "NON_PARTICIPATING", "128": "NON_PARTICIPATING", "129": "NOT_ALIVE", "13": "NOT_ALIVE", "130": "NOT_ALIVE", "131": "NOT_ALIVE", "132": "NOT_ALIVE", "133": "SUCCESS", "134": "NON_PARTICIPATING", "135": "NOT_ALIVE", "136": "NON_PARTICIPATING", "137": "NON_PARTICIPATING", "138": "NON_PARTICIPATING", "139": "NON_PARTICIPATING", "14": "NOT_ALIVE", "140": "NON_PARTICIPATING", "141": "NOT_ALIVE", "142": "NOT_ALIVE", "143": "NON_PARTICIPATING", "144": "NOT_ALIVE", "145": "NON_PARTICIPATING", "146": "NON_PARTICIPATING", "147": "NOT_ALIVE", "148": "NON_PARTICIPATING", "149": "NON_PARTICIPATING", "15": "NOT_ALIVE", "150": "NOT_ALIVE", "151": "NON_PARTICIPATING", "152": "NOT_ALIVE", "153": "NON_PARTICIPATING", "154": "NOT_ALIVE", "155": "NOT_ALIVE", "156": "NOT_ALIVE", "157": "NOT_ALIVE", "158": "NON_PARTICIPATING", "159": "NOT_ALIVE", "16": "NOT_ALIVE", "160": "NON_PARTICIPATING", "161": "NON_PARTICIPATING", "162": "NOT_ALIVE", "163": "NON_PARTICIPATING", "164": "NOT_ALIVE", "165": "NON_PARTICIPATING", "166": "SUCCESS", "167": "NON_PARTICIPATING", "168": "NON_PARTICIPATING", "169": "NOT_ALIVE", "17": "NOT_ALIVE", "170": "NON_PARTICIPATING", "171": "NOT_ALIVE", "172": "NON_PARTICIPATING", "173": "NON_PARTICIPATING", "174": "NOT_ALIVE", "175": "NOT_ALIVE", "176": "NOT_ALIVE", "177": "NOT_ALIVE", "178": "NON_PARTICIPATING", "179": "NON_PARTICIPATING", "18": "NOT_ALIVE", "180": "NON_PARTICIPATING", "181": "NON_PARTICIPATING", "182": "NON_PARTICIPATING", "183": "NOT_ALIVE", "184": "NOT_ALIVE", "185": "NON_PARTICIPATING", "186": "NOT_ALIVE", "187": "SUCCESS", "188": "NOT_ALIVE", "189": "NON_PARTICIPATING", "19": "NOT_ALIVE", "190": "SUCCESS", "191": "NON_PARTICIPATING", "192": "NOT_ALIVE", "193": "NOT_ALIVE", "194": "NON_PARTICIPATING", "195": "NON_PARTICIPATING", "196": "NOT_ALIVE", "197": "NON_PARTICIPATING", "198": "NON_PARTICIPATING", "199": "NOT_ALIVE", "2": "NOT_ALIVE", "20": "NOT_ALIVE", "200": "NOT_ALIVE", "201": "SUCCESS", "202": "NON_PARTICIPATING", "203": "NOT_ALIVE", "204": "NOT_ALIVE", "205": "NOT_ALIVE", "206": "NON_PARTICIPATING", "207": "NOT_ALIVE", "208": "NOT_ALIVE", "209": "NOT_ALIVE", "21": "NOT_ALIVE", "210": "NON_PARTICIPATING", "211": "NOT_ALIVE", "212": "NON_PARTICIPATING", "213": "NON_PARTICIPATING", "214": "NON_PARTICIPATING", "215": "NON_PARTICIPATING", "216": "NON_PARTICIPATING", "217": "NON_PARTICIPATING", "218": "NON_PARTICIPATING", "219": "SUCCESS", "22": "NOT_ALIVE", "220": "NOT_ALIVE", "221": "NOT_ALIVE", "222": "NOT_ALIVE", "223": "NON_PARTICIPATING", "224": "NON_PARTICIPATING", "225": "NON_PARTICIPATING", "226": "NON_PARTICIPATING", "227": "NOT_ALIVE", "228": "SUCCESS", "229": "NOT_ALIVE", "23": "NOT_ALIVE", "230": "NON_PARTICIPATING", "231": "NOT_ALIVE", "232": "NOT_ALIVE", "233": "NOT_ALIVE", "234": "NON_PARTICIPATING", "235": "NOT_ALIVE", "236": "NOT_ALIVE", "237": "NON_PARTICIPATING", "238": "NON_PARTICIPATING", "239": "NOT_ALIVE", "24": "NOT_ALIVE", "240": "NOT_ALIVE", "241": "NON_PARTICIPATING", "242": "NON_PARTICIPATING", "243": "NOT_ALIVE", "244": "NON_PARTICIPATING", "245": "NOT_ALIVE", "246": "NON_PARTICIPATING", "247": "NON_PARTICIPATING", "248": "NOT_ALIVE", "249": "NOT_ALIVE", "25": "SUCCESS", "250": "NON_PARTICIPATING", "251": "NOT_ALIVE", "252": "NON_PARTICIPATING", "253": "NON_PARTICIPATING", "254": "NOT_ALIVE", "255": "NON_PARTICIPATING", "26": "NOT_ALIVE", "27": "NOT_ALIVE", "28": "NON_PARTICIPATING", "29": "NON_PARTICIPATING", "3": "SUCCESS", "30": "NOT_ALIVE", "31": "SUCCESS", "32": "NON_PARTICIPATING", "33": "NOT_ALIVE", "34": "NON_PARTICIPATING", "35": "NOT_ALIVE", "36": "NOT_ALIVE", "37": "NON_PARTICIPATING", "38": "NOT_ALIVE", "39": "NOT_ALIVE", "4": "NOT_ALIVE", "40": "NON_PARTICIPATING", "41": "NOT_ALIVE", "42": "NON_PARTICIPATING", "43": "NOT_ALIVE", "44": "NOT_ALIVE", "45": "NON_PARTICIPATING", "46": "NON_PARTICIPATING", "47": "NOT_ALIVE", "48": "NOT_ALIVE", "49": "NON_PARTICIPATING", "5": "NOT_ALIVE", "50": "NOT_ALIVE", "51": "NOT_ALIVE", "52": "NON_PARTICIPATING", "53": "NON_PARTICIPATING", "54": "NOT_ALIVE", "55": "NON_PARTICIPATING", "56": "NON_PARTICIPATING", "57": "NON_PARTICIPATING", "58": "NOT_ALIVE", "59": "NOT_ALIVE", "6": "NOT_ALIVE", "60": "NON_PARTICIPATING", "61": "NON_PARTICIPATING", "62": "NOT_ALIVE", "63": "NON_PARTICIPATING", "64": "NON_PARTICIPATING", "65": "NON_PARTICIPATING", "66": "NON_PARTICIPATING", "67": "NOT_ALIVE", "68": "NOT_ALIVE", "69": "NON_PARTICIPATING", "7": "NOT_ALIVE", "70": "NON_PARTICIPATING", "71": "NON_PARTICIPATING", "72": "NOT_ALIVE", "73": "NON_PARTICIPATING", "74": "NON_PARTICIPATING", "75": "NON_PARTICIPATING", "76": "NON_PARTICIPATING", "77": "NON_PARTICIPATING", "78": "NON_PARTICIPATING", "79": "NOT_ALIVE", "8": "NOT_ALIVE", "80": "NOT_ALIVE", "81": "NOT_ALIVE", "82": "NON_PARTICIPATING", "83": "SUCCESS", "84": "NON_PARTICIPATING", "85": "NOT_ALIVE", "86": "NON_PARTICIPATING", "87": "NOT_ALIVE", "88": "NOT_ALIVE", "89": "NOT_ALIVE", "9": "NOT_ALIVE", "90": "NOT_ALIVE", "91": "NOT_ALIVE", "92": "NON_PARTICIPATING", "93": "NON_PARTICIPATING", "94": "NON_PARTICIPATING", "95": "NON_PARTICIPATING", "96": "NON_PARTICIPATING", "97": "NOT_ALIVE", "98": "NOT_ALIVE", "99": "SUCCESS" }, "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "block_list": [], "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "inner_step": 0, "intermediate_size": 5632, "last_allreduce_block": 6168147, "max_position_embeddings": 2048, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 22, "num_key_value_heads": 4, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 10000.0, "tie_word_embeddings": false, "torch_dtype": "float32", "transformers_version": "4.39.3", "use_cache": false, "vocab_size": 32000 }