{ "_name_or_path": "distributed/llama-1b", "all_reduce_scores": { "0": "NOT_ALIVE", "1": "NON_PARTICIPATING", "10": "NON_PARTICIPATING", "100": "NOT_ALIVE", "101": "NOT_ALIVE", "102": "NON_PARTICIPATING", "103": "NON_PARTICIPATING", "104": "NON_PARTICIPATING", "105": "NON_PARTICIPATING", "106": "NON_PARTICIPATING", "107": "NOT_ALIVE", "108": "NOT_ALIVE", "109": "NOT_ALIVE", "11": "NON_PARTICIPATING", "110": "NON_PARTICIPATING", "111": "NON_PARTICIPATING", "112": "NOT_ALIVE", "113": "NOT_ALIVE", "114": "NON_PARTICIPATING", "115": "NON_PARTICIPATING", "116": "NOT_ALIVE", "117": "NON_PARTICIPATING", "118": "NON_PARTICIPATING", "119": "NON_PARTICIPATING", "12": "NON_PARTICIPATING", "120": "NON_PARTICIPATING", "121": "NOT_ALIVE", "122": "NOT_ALIVE", "123": "NON_PARTICIPATING", "124": "NON_PARTICIPATING", "125": "NON_PARTICIPATING", "126": "NOT_ALIVE", "127": "NON_PARTICIPATING", "128": "NOT_ALIVE", "129": "NON_PARTICIPATING", "13": "NON_PARTICIPATING", "130": "NON_PARTICIPATING", "131": "NOT_ALIVE", "132": "NON_PARTICIPATING", "133": "NON_PARTICIPATING", "134": "NON_PARTICIPATING", "135": "NON_PARTICIPATING", "136": "NON_PARTICIPATING", "137": "NON_PARTICIPATING", "138": "NON_PARTICIPATING", "139": "NOT_ALIVE", "14": "SUCCESS", "140": "NOT_ALIVE", "141": "NOT_ALIVE", "142": "NON_PARTICIPATING", "143": "NON_PARTICIPATING", "144": "NOT_ALIVE", "145": "NON_PARTICIPATING", "146": "NON_PARTICIPATING", "147": "NOT_ALIVE", "148": "NON_PARTICIPATING", "149": "NON_PARTICIPATING", "15": "NON_PARTICIPATING", "150": "NON_PARTICIPATING", "151": "NOT_ALIVE", "152": "NOT_ALIVE", "153": "NON_PARTICIPATING", "154": "NON_PARTICIPATING", "155": "NON_PARTICIPATING", "156": "NON_PARTICIPATING", "157": "NON_PARTICIPATING", "158": "NOT_ALIVE", "159": "NON_PARTICIPATING", "16": "NOT_ALIVE", "160": "NON_PARTICIPATING", "161": "NON_PARTICIPATING", "162": "NON_PARTICIPATING", "163": "NON_PARTICIPATING", "164": "SUCCESS", "165": "NON_PARTICIPATING", "166": "NON_PARTICIPATING", "167": "NON_PARTICIPATING", "168": "NON_PARTICIPATING", "169": "NOT_ALIVE", "17": "NON_PARTICIPATING", "170": "NON_PARTICIPATING", "171": "NOT_ALIVE", "172": "NOT_ALIVE", "173": "NON_PARTICIPATING", "174": "NON_PARTICIPATING", "175": "NOT_ALIVE", "176": "NOT_ALIVE", "177": "NOT_ALIVE", "178": "NON_PARTICIPATING", "179": "NON_PARTICIPATING", "18": "NOT_ALIVE", "180": "NOT_ALIVE", "181": "NOT_ALIVE", "182": "NOT_ALIVE", "183": "SUCCESS", "184": "NON_PARTICIPATING", "185": "NON_PARTICIPATING", "186": "NON_PARTICIPATING", "187": "NOT_ALIVE", "188": "NOT_ALIVE", "189": "NOT_ALIVE", "19": "NOT_ALIVE", "190": "NON_PARTICIPATING", "191": "NOT_ALIVE", "192": "NON_PARTICIPATING", "193": "NON_PARTICIPATING", "194": "NON_PARTICIPATING", "195": "NON_PARTICIPATING", "196": "NON_PARTICIPATING", "197": "NOT_ALIVE", "198": "NOT_ALIVE", "199": "NON_PARTICIPATING", "2": "NON_PARTICIPATING", "20": "NON_PARTICIPATING", "200": "NOT_ALIVE", "201": "NOT_ALIVE", "202": "NON_PARTICIPATING", "203": "NON_PARTICIPATING", "204": "NON_PARTICIPATING", "205": "NON_PARTICIPATING", "206": "NON_PARTICIPATING", "207": "NOT_ALIVE", "208": "NOT_ALIVE", "209": "NOT_ALIVE", "21": "NOT_ALIVE", "210": "NOT_ALIVE", "211": "NON_PARTICIPATING", "212": "NON_PARTICIPATING", "213": "NON_PARTICIPATING", "214": "NON_PARTICIPATING", "215": "SUCCESS", "216": "NOT_ALIVE", "217": "NON_PARTICIPATING", "218": "NON_PARTICIPATING", "219": "NON_PARTICIPATING", "22": "NON_PARTICIPATING", "220": "NON_PARTICIPATING", "221": "NOT_ALIVE", "222": "NOT_ALIVE", "223": "NON_PARTICIPATING", "224": "NOT_ALIVE", "225": "NON_PARTICIPATING", "226": "NOT_ALIVE", "227": "NON_PARTICIPATING", "228": "NON_PARTICIPATING", "229": "NON_PARTICIPATING", "23": "NON_PARTICIPATING", "230": "NON_PARTICIPATING", "231": "NOT_ALIVE", "232": "NOT_ALIVE", "233": "NON_PARTICIPATING", "234": "NON_PARTICIPATING", "235": "NON_PARTICIPATING", "236": "NOT_ALIVE", "237": "NON_PARTICIPATING", "238": "NON_PARTICIPATING", "239": "NON_PARTICIPATING", "24": "NON_PARTICIPATING", "240": "NOT_ALIVE", "241": "NON_PARTICIPATING", "242": "NON_PARTICIPATING", "243": "NOT_ALIVE", "244": "NON_PARTICIPATING", "245": "SUCCESS", "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": "NON_PARTICIPATING", "255": "NOT_ALIVE", "26": "NON_PARTICIPATING", "27": "NON_PARTICIPATING", "28": "NOT_ALIVE", "29": "NON_PARTICIPATING", "3": "NON_PARTICIPATING", "30": "SUCCESS", "31": "NON_PARTICIPATING", "32": "NON_PARTICIPATING", "33": "NOT_ALIVE", "34": "NOT_ALIVE", "35": "NOT_ALIVE", "36": "SUCCESS", "37": "NOT_ALIVE", "38": "SUCCESS", "39": "SUCCESS", "4": "NOT_ALIVE", "40": "NON_PARTICIPATING", "41": "NOT_ALIVE", "42": "SUCCESS", "43": "NON_PARTICIPATING", "44": "NON_PARTICIPATING", "45": "NOT_ALIVE", "46": "NOT_ALIVE", "47": "NON_PARTICIPATING", "48": "NOT_ALIVE", "49": "NON_PARTICIPATING", "5": "SUCCESS", "50": "NON_PARTICIPATING", "51": "NON_PARTICIPATING", "52": "NON_PARTICIPATING", "53": "NON_PARTICIPATING", "54": "NOT_ALIVE", "55": "NOT_ALIVE", "56": "SUCCESS", "57": "NON_PARTICIPATING", "58": "NON_PARTICIPATING", "59": "SUCCESS", "6": "NON_PARTICIPATING", "60": "NOT_ALIVE", "61": "NOT_ALIVE", "62": "SUCCESS", "63": "SUCCESS", "64": "NON_PARTICIPATING", "65": "NON_PARTICIPATING", "66": "NON_PARTICIPATING", "67": "NOT_ALIVE", "68": "NON_PARTICIPATING", "69": "NON_PARTICIPATING", "7": "SUCCESS", "70": "NON_PARTICIPATING", "71": "NON_PARTICIPATING", "72": "NON_PARTICIPATING", "73": "NON_PARTICIPATING", "74": "NOT_ALIVE", "75": "NOT_ALIVE", "76": "NON_PARTICIPATING", "77": "NON_PARTICIPATING", "78": "NON_PARTICIPATING", "79": "NON_PARTICIPATING", "8": "NON_PARTICIPATING", "80": "NOT_ALIVE", "81": "NON_PARTICIPATING", "82": "NOT_ALIVE", "83": "NON_PARTICIPATING", "84": "NOT_ALIVE", "85": "NON_PARTICIPATING", "86": "NON_PARTICIPATING", "87": "NON_PARTICIPATING", "88": "NON_PARTICIPATING", "89": "NOT_ALIVE", "9": "NOT_ALIVE", "90": "NON_PARTICIPATING", "91": "NOT_ALIVE", "92": "NOT_ALIVE", "93": "NOT_ALIVE", "94": "NON_PARTICIPATING", "95": "NOT_ALIVE", "96": "NOT_ALIVE", "97": "NON_PARTICIPATING", "98": "NOT_ALIVE", "99": "NOT_ALIVE" }, "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "block_list": [ 5950315, 5950318, 5950322 ], "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "inner_step": 237, "intermediate_size": 5632, "last_allreduce_block": 5948952, "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 }